U.S. JGOFS Synthesis & Modeling Project
Principal Investigator's Workshop
July 22-26, 2002

back to Agenda
Anderson, L et al. Anderson, R* Armstrong & Landry Bauer et al. Bogucki & Redekopp
Buesseler Carr et al. Chai et al. #1 Chai et al. #2 Christian & Letelier
Christian & Murtugudde Chung et al. Daniels et al. DeMaster et al. Deutsch et al.
Dinniman et al. Ducklow et al.* Dunne et al. Falkowski* Feely et al.
Fennel et al. Francois* Friedrichs et al. Glover & Conte Gruber et al. #1*
Gruber et al. #2 Gruber et al. #3 Hood et al. Irwin & Falkowski Irwin et al.
Ito et al. Jiang et al. Karl* Lampitt et al. Laws
Lee et al. Leising et al. Lima & Doney Lima et al. Litchman et al.
Matsumoto et al. McGillicuddy et al. McKinley et al. McNeil et al. Mishonov et al.
Mongin & Nelson #1 Mongin & Nelson #2 Moore* Moore et al. Murtuggude & Christian
Parekh et al. Peng & Li Prézelin et al. Redalje et al. Richardson MJ et al.
Richardson T et al. Robbins Sabine et al. #1* Sabine et al. #2 Salihoglu & Hofmann
Sarmiento & Dunne* Schlitzer* Smith et al. Spitz Sweeney
Toole et al. Turk Tréguer Wiggert et al. Yoder

* Plenary Speaker

Larry Anderson1, Dennis McGillicuddy1, Scott Doney1, Mat Maltrud2 and Frank Bryan3

The Role of Mesoscale Eddies in Basin-scale Biogeochemical Budgets of the North Atlantic

1 Woods Hole Oceanographic Institution, Woods Hole, MA 02543
2 Los Alamos National Laboratory, Los Alamos, NM 87545
3 National Center for Atmospheric Research, Boulder, CO 80307

Four 4-year eddy-resolving (0.1-degree) simulations of nitrate transport in the euphotic zone of the North and Equatorial Atlantic have been conducted and are compared with data, coarse-resolution simulations and previous modeling studies. Through the lifting of nitrate-rich isopycnals into the euphotic zone, vertical advection by mesoscale eddies dominates the transport of nitrate into the euphotic zone of the permanently-stratified subtropical gyre. Coarse-resolution simulations run with the Gent and McWilliams (1990) mesoscale isopycnal mixing parameterization do not capture this effectively diapycnal flux.

Robert Anderson1

The seasonal cycle of the Antarctic Circumpolar Current at 170°W - Ecosystem structure, nutrient utilization and export flux.

1 Lamont-Doherty Earth Observatory, Palisades, NY 10964

Among the three principal High-Nutrient Low-Chlorophyll regions (Subarctic Pacific, Equatorial Pacific and Southern Ocean), the Southern Ocean contains by far the largest inventory of surface nutrients. The biological pump, measured in terms of the fraction of upwelled nutrients converted to organic matter and exported to the deep sea, is clearly not high in the Southern Ocean. Nutrient utilization efficiency in the Southern Ocean is potentially an important factor within the array of mechanisms by which ocean processes regulate the CO2 content of the atmosphere. Consequently, much attention has been paid to this topic, and many investigators have speculated about the relative importance of the factors (light, grazing, iron) that potentially limit the growth of phytoplankton and the consumption of nutrients in the Southern Ocean.

Studies conducted by US JGOFS in the SW Pacific sector of the Southern Ocean provide an unprecedented view of the seasonal evolution of the structure of the planktonic ecosystem, as well as of factors regulating nutrient utilization and export of biogenic material to depth. Nutrient utilization and export flux is intimately linked to the growth of large diatom species, although diatoms are not the sole agent of export. At various places, and at different times, light, grazing, iron and dissolved silicic acid each play a limiting role. These complex relationships must all be considered when simulating the sensitivity of the Southern Ocean's biological pump to climate change.

Rob Armstrong1 and Mike Landry2

A model structure for analysis of food web data and for carbon cycle simulation

1 Marine Science Res. Center, Stony Brook Univ., Stony Brook, NY
2 Dept of Oceanography, Univ. of Hawaii, Honolulu, HI

Previous size-structured models of marine food webs have proved inadequate for analyzing food web data. In particular, the use of discrete, non-overlapping "boxes" for size classes does not capture the size plasticity of individual organisms, nor does it reflect the fact that characteristic size differences among competing species may not be the same as characteristic size differences between adjacent trophic levels. Here we present a new food web model, with a parameterized zooplankton size spectrum, that we believe will enable maximal information to be extracted from food web data. This model structure should also be useful in simulating ecosystem processes in food web models. Development of this model was part of our ongoing collaborative project to model food web transfers and export fluxes in pelagic ecosystems.

Armstrong, R.A. Beyond Moloney and Field: A hybrid spectral model of plankton interaction. DSR II (submitted)

J.E. Bauer1, M.D. DeGrandpre2, P. Vlahos3, C.S. Hopkinson4, R.F. Chen5 and L.I. Aluwihare6.

Whole-Shelf and Slope Studies of Carbon Inventories and Fluxes in the Northwest Atlantic Continental Margin During the Ocean Margins Program

1 School of Marine Science/VIMS, College of Wm. & Mary, Gloucester Pt., VA 23062
2 University of Montana, Dept. of Chemistry, 32 Campus Dr., Missoula, MT 59812
3 University of Connecticut, Dept. of Marine Sciences,1084 Shennecossett Rd., Groton, CT 06340
4 Ecosystems Center, Marine Biological Laboratory, Woods Hole, MA 02543
5 University of Massachusetts, Dept. of Environmental, Coastal and Ocean Sciences, 100 Morrissey Blvd., Boston, MA 02125
6 Scripps Institute of Oceanography, Geosciences Research Division, 9500 Gilman Dr., La Jolla, CA 92093

The DOE- and NSF-funded Ocean Margins Program undertook a comprehensive evaluation of the inventories (including sources and inputs) and fluxes of organic and inorganic carbon in shelf and slope waters of the Middle Atlantic Bight (MAB) in the NW Atlantic from 1994-1996. These studies included both whole-system shipboard surveys from Cape Cod to Cape Hatteras, as well as autonomous measurements from moored arrays in the southern portion of the study area. In spite of a long history of carbon-based studies in this region, there have been no previously reported measurements of either the partial pressure of CO2 (pCO2) or of dissolved organic carbon (DOC) for the MAB prior to 1994, thus precluding the synthesis of a comprehensive carbon budget there. Calculations based on air-sea pCO2 differences indicate that the MAB is a net annual sink for atmospheric CO2 with the inner, mid, and outer-shelf regions taking up ~0.1, 0.7, and 0.2 Mt C yr-1, respectively, for a net uptake of ~1&plusnm;0.6 Mt C yr-1, which is relatively small compared to the net influxes observed recently in some other ocean margins. The annual cycle of heating and cooling combined with high winds during the period of undersaturation (winter) appears to account for a significant portion of the uptake. The flux uncertainty is dominated by uncertainty of the gas transfer velocity parameterization, atmospheric CO2 levels, and coarse spatial pCO2 resolution.

Concentrations of DOC were greater inshore than offshore and increased southward along the shelf. The total DOC inventory on the shelf during spring 1996 was estimated at ~5.88 x 1012 g C and this increased by 0.4 x 1012 g C (7%) by late summer. A simple mass balance of DOC input and export in the MAB resulted in total export of 18.7 to 19.6 x 1012 g C y-1. Although water budgets for the MAB suggest a relatively small (~5-10%) input flux of DOC from land, a dual isotopic (14C and 13C) approach and multi-source model suggests that as much as half of the total bulk DOC, and even greater fraction of the suspended POC, originates from terrigenous sources that have become highly aged either on land or in the MAB. On average, the C:N:P ratio of shelf DOM (431:36:1) was substantially higher than Redfield, but not nearly as high as that of deep slope water (2700:215:1); the selective bacterial remineralization of P and N from DOM was further confirmed by incubation studies. The presence of a young, bomb 14C-enriched polysaccharide fraction was observed in the >1kDalton DOM fraction, and may be representative of the more labile and semi-labile components of DOM in this system.

D. Bogucki1 and L. G. Redekopp2

Absolute and convective instabilities in air-water interface


Current estimates of air-sea gas exchange vary by over a factor of two depending on the parameterization of the transfer velocity from wind speed. Uncertainties in the parameterization occur because wind speed is indicative of, but not necessarily an accurate descriptor for, the local characteristics of the air-water interface in the presence of both the air and water boundary layers.

To obtain greater insight to the dynamics of the air-water boundary layer, we are investigating a generalized Holmboe model of the air-water interface. We analytically investigate the onset of instabilities considering both spatial and temporal modes, with the goal of identifying conditions for spontaneous onset of intrinsic dynamics that are not dependent on the spatial history of the flow. The results will be compared to observations from tank experiments in the future.

Ken O. Buesseler1

Magnitude, Variability and Controls on the Ratio of Particle Export to Primary Production in the Upper Ocean

1 Woods Hole Oceanographic Institution Woods Hole, MA 02543 USA tel: 508-289-2309; fax: 508-457-2193 email: kbuesseler@whoi.edu

The transport of biogenic particles from the surface to the deep ocean is the key driver of the ocean's biological pump. Globally, the magnitude and efficiency of the biological pump will in part modulate levels of atmospheric CO2, and from the geological paleo-oceanographic record there is evidence of elevated rates of export of POC resulting from changes in the functioning of the pump. Thus there is a need to better understand what are the key determinants of this pump in the present day, and how they might be altered in response to climate change. This SMP project and poster examine the present day relationship between primary production and particulate export in the upper ocean. Recent advances in satellite derived algorithms for primary production lend well to improved global predictions of the rate of C uptake, however our ability to determine particle fluxes is much poorer. A pronounced mismatch between spatial patterns in primary production and the export of carbon to the deep ocean, points to the complex suite of transformations that occur in the upper 300 m of the ocean. The results thus far indicate that the relative rates of C uptake and losses via sinking particles vary as a function of the local food web dynamics. In particular, diatoms appear to play an important role in enhancing the ratio of export:production in the upper ocean. In spite of the recent development of promising modeling approaches to assess export production on global scales (Laws et al., 2000), our understanding of the key processes determining what controls the efficiency of particle transport between the surface and deep ocean remains weak. Results from a recent synthesis of the AESOPS data along the Polar Front will be highlighted in this poster, as well as an update on the compilation of a global shallow POC flux budget as part of this SMP project.

Mary-Elena Carr1, Marjorie Friedrichs2 and Ahmed H. Ali1

Marine primary production estimates from ocean color: a comparative study of algorithms

1 Jet Propulsion Laboratory, California Institute of Technology
2 Old Dominion Unversity

The Primary Production Algorithm Round Robin 3 (PPARR3) aims to compare models or algorithms that estimate marine primary production from satellite measurements of ocean color (PP models). It is a continuation of previous PPARR exercises, which compared in situ carbon14 uptake rates with an estimate of primary production using satellite-accessible data. PPARR2 found that modeled primary production would be within a factor of two of the in situ rates if systematic offsets were corrected. PPARR3 aims to provide a forum to compare model output, improve parameterization, and help identify the source of biases. This community project presently counts with over twenty modeling groups who estimate primary production for input fields provided by the organizers. The PPARR3 exercise consists of 3 stages, the first stage is a comparison of monthly global primary production fields generated by the different algorithms. Stage 2 is a step-by-step sensitivity study of the different algorithms tracking the derivation of sub-products in a series of point value estimates. The third stage is similar to PPARR1 and PPARR2 and is a blind comparison to the quality-controlled data base of carbon-14 measurements in the equatorial Pacific. We present here the results of the first stage, which compares the output of the models throughout an annual cycle.

Takamitsu Ito1, Mick Follows1,and John Marshall1

An idealized model of tracer transport and ventilation in the Southern Ocean.

1 Program in Atmospheres Oceans and Climate, Dept Earth, Atmospheric and Planetary Sciences, Massachusetts Inst. of Technology, Cambridge, MA 02139

We construct a zonally-averaged model of tracer transport and air-sea gas exchanges in the Southern Ocean based on the residual mean theory. Given the surface wind stress and the surface buoyancy forcing, the idealized theory can predict the interior stratification and the meridional overturning circulation (MOC) in the vicinity of the Antarctic Circumpolar Current (ACC). We evaluate the simple theory by calculating the time-varying, zonally-averaged distribution of CFCs in the region of the ACC and comparing to observations from the WOCE survey. We find that the simple theory captures the broad structure of ventilation of CFCs into the thermocline and intermediate waters very well, and can be extended to represent the formation of deep waters, produced by the insense buoyancy fluxes near Antarctica. The simplicity of the model allows a clear examination of the relationships between the physical forcing, eddy transfer mechanisms and tracer transports. We discuss the sensitivities of the ventilation of CFCs to the imposed surface wind stress and the buoyancy forcing. We find that the ventilation of CFCs into the thermocline and intermediate waters is controlled both by advection in the residual circulation and isopycnal eddy stirring. The vigor of the residual circulation and MOC is most sensitive to the amplitiude and patterns of surface buoyancy fluxes while the isopycnal stirring of tracers is sensitive to the surface wind stress. This idealized model provides insights into the mechanisms controlling the ventilation of tracers into the oceans and in more complex, three-dimensional general circulation models and can be extended to consider biogeochemically active tracers.

F. Chai1, M.-S. Jiang1, R. T. Barber2, R. C. Dugdale3 and Y. Chao4

Interdecadal Variation of the Transition Zone Chlorophyll Front, A Physical-Biological Model Simulation between 1960 and 1990

1 School of Marine Sciences, University of Maine, Orono, ME 04469-5741
2 Duke University, NSOE Marine Laboratory, 135 Duke Marine Lab Road, Beaufort, NC 28516
3 Romberg Tiburon Center, San Francisco State University, PO Box 855, Tiburon CA 94920
4 Jet Propulsion Laboratory, 4800 Oak Grove Drive, Pasadena, CA 91109

The Transition Zone Chlorophyll Front (TZCF) separates the low chlorophyll subtropical gyres and the high chlorophyll subarctic gyres in the Pacific Ocean. The interdecadal climate variability affects marine ecosystems in both subtropical and subarctic gyres, consequently the position of the TZCF. A three-dimensional physical-biological model has been used to study interdecadal variation of the TZCF using a retrospective analysis of a 30-year (1960-1990) model simulation. The physical-biological model is forced with the monthly mean heat flux and surface wind stress from the Comprehensive Ocean Atmosphere Data Set.

The modeled position of the TZCF, operationally defined as the isopleth of 0.2 mg/m3 chlorophyll, is located between 25°N and 27°N in the central North Pacific during the winter and between 33°N and 35°N during the summer, which agrees with the seasonal migration patterns of the TZCF detected with SeaWiFS. The modeled winter MLD shows the largest increase between 30°N and 40°N in the central North Pacific (150°E to 180°), with a value of 40-60% higher (deeper mixed layer) during 1979-90 relative to 1964-75 values. In the subarctic gyre in both northeast (Ocean Station Papa, OSP) and northwest Pacific (Oyashio region), the modeled winter MLD decreases by about 20% during the period of 1979-90 relative to 1964-75 levels. The winter Ekman pumping velocity difference between 1979-90 and 1964-75 shows the largest increase is located between 30°N and 45°N in the central and eastern North Pacific (180 to 150°W). In the subarctic northeast Pacific region including the Gulf of Alaska, the winter Ekman pumping velocity decreases during the period of 1979-90, but its value increases in the northwest Pacific (Oyashio region) after 1976-77 climatic shift. The modeled winter surface nitrate difference between 1979-90 and 1964-75 shows increase in the latitudinal band of 30°N and 45°N from the west to the east (135°E-135°W), the modeled nitrate concentration is about 10 to 50% higher in general during the period of 1979-90 relative to 1964-75 values depending upon location. The increase of the winter surface nitrate concentration during 1979-90 is caused by a combination of the winter MLD increase and the winter Ekman pumping enhancement after 1976-77 climatic shift. The modeled nitrate concentration increase after 1976-77 lead to the primary productivity increase in the central North Pacific (30°N-40°N and 180°-140°W). Enhanced primary productivity after the 1976-77 climatic shift contributes higher phytoplankton biomass and therefore elevates chlorophyll level in the central North Pacific. Increase in the modeled chlorophyll expend the transitional zone and push the TZCF equatorward.

F. Chai1, M.-S. Jiang1, R. T. Barber2, R. A. Feely3, R. C. Dugdale4, T.-H. Peng5 and Y. Chao6

Modeled Decadal Variability of Primary Productivity and Air-Sea CO2 Flux in the Equatorial Pacific Ocean

1 School of Marine Science, 5471 Libby Hall, University of Maine, Orono, ME 04469
2 Duke University, NSOE Marine Laboratory, 135 Duke Marine Lab Road, Beaufort, NC 28516
3 Richard A. Feely (Ocean Climate Research Division, NOAA/PMEL, 7600 Sand Point Way NE, Seattle WA 98115
4 Romberg Tiburon Center, San Francisco State University, PO Box 855, Tiburon CA 94920
5 NOAA Atlantic Oceanographic and Meteorological Laboratory, Ocean Chemistry Division, 4301 Rickenbacker Causeway, Miami, FL 33149-1026
6 Jet Propulsion Laboratory, California Institute of Technology, 4800 Oak Grove Drive, Pasadena, CA 91109

The response of primary production and sea-to-air CO2 flux in the equatorial Pacific to decadal timescale climate variability is investigated using a physical-biogeochemical model forced with COADS wind stress and heat flux. The circulation model resolves decadal variations over the past 40 years with a decrease in the equatorward interior flow and a reduction of 20% in the equatorial upwelling transport since 1976-77 climate shift. The decreased volume transports causes a rise of sea surface temperature in the equatorial upwelling zone by about 0.7°C since mid 1970s. Slowdown of the meridional overturning and decrease of the equatorial upwelling transport have significant impacts on marine ecosystem and carbon flux. The modeled primary production and phytoplankton biomass decrease by 10% over the past 40 years mainly due to reduction of upward nutrient flux. Our physical-biogeochemical model results document that the equatorial Pacific sea-to-air CO2 flux decreased by 20% after 1976-77 climate shift.

James Christian1 and Ricardo Letelier2

Modelling interannual variability of carbon fluxes at JGOFS Time-Series Station ALOHA

1 Earth System Science Interdisciplinary Center, University of Maryland
2 College of Oceanic and Atmospheric Sciences, Oregon State University

The Hawaii Ocean Time-Series has collected one of the most complete data sets on the temporal variability of dissolved and particulate carbon, nitrogen and phosphorus for any location in the world ocean. To date few prognostic ecosystem models have been coupled to models of carbon chemistry and air-sea exchange, and most have considered only C and N, and have employed fixed elemental (Redfield) ratios. We have developed models for C-N-P stoichiometry that consider both inorganic and organic dissolved pools, with fixed and variable ratios in phytoplankton biomass; regeneration terms are common to both models. The results of these models show that the biologically mediated oceanic sink for atmospheric CO2 is consistently larger in the variable-ratio model. Interannual variability of surface ocean dissolved inorganic carbon (DIC) is large and is forced primarily by variability of precipitation. Covariance of modelled and observed surface salinity and DIC suggest that the precipitation fields generated by the Global Precipitation Climatology Project are quite accurate at this location except during 1994-1995 when precipitation appears to be substantially underestimated. The secular trend in salinity-normalized DIC for 1988-2000 is 0.88 mmol m-3 y-1 in the upper 50 m, consistent with observation-based estimates. This trend is primarily caused by rising atmospheric CO2 although up to 25% may be attributable to interannual variability in entrainment of subsurface water. The range of DIC variability associated with variable freshwater flux is about 100 times the annual accumulation, and there is an apparent secular trend in total DIC that is about twice as large as the secular trend in normalized DIC. This trend is due principally to declining precipitation after 1998.

James Christian1 and Ragu Murtugudde1

Tropical Atlantic climate variability in a coupled physical-biogeochemical ocean model

1 Earth System Science Interdisciplinary Center, University of Maryland, College Park, College Park, MD, 20742

A three-dimensional ocean biogeochemical model of the tropical Atlantic Ocean was run for more than half a century (1949-2000) in order to characterize the ocean biogeochemical response to variable forcing over this period. The seasonal cycle in the equatorial upwelling zone agrees reasonably well with observations and other published simulations but underestimates phytoplankton biomass under strong upwelling conditions. Away from the equator, nutrient flux and biological production are maximal in each hemisphere's winter season, and appear to be proximately forced by evapourative cooling and wind stirring rather than by Ekman upwelling. The fraction of the total variance that is associated with the annual cycle is considerably smaller for biogeochemical fields than for sea surface temperature over this long simulation, and much of this variance is associated with interdecadal changes. The tropical Atlantic appears to have become more productive following the Pacific climate shift of 1976 and remained so until about 1989. Summer surface nitrate concentrations during the 1990's were lower than those in the 1980's. The relationship between the equatorial and off-equatorial regimes may have changed following the 1976 event, with equatorial variability dominating the basin-wide variance patterns after 1976.

S. Chung1, Kitack Lee1, C.L. Sabine2,3, R.A. Feely3, F.J. Millero4, R.M. Key5, and R. Wanninkhof6

Calcium Carbonate Budget in the Atlantic Ocean

1 SEE, Pohang University of Science and Technology, Pohang, Korea, email: ktl@postech.ac.kr
2 JISAO, University of Washington, Seattle
3 Pacific Marine Environmental Laboratory, NOAA, Seattle
4 RSMAS, University of Miami, Miami
5 AOSP, Princeton University, Princeton, New Jersey
6 Atlantic Oceanographic and Meteorological Laboratory, NOAA, Miami

Several independent lines of recent evidence suggest that the dissolution of calcium carbonate (CaCO3) particles is substantial in the upper ocean above the 100% calcite saturation horizon. However, the direct evidence is not available and an underlying mechanism remains unresolved. This shallow-water dissolution of carbonate particles is in conflict with the long-held paradigm of the conservative nature of pelagic CaCO3 at shallow water depths. Here we used over 20,000 carbon measurements in conjunction with CFC data from the WOCE/JGOFS global CO2 survey to estimate the in-situ dissolution rates of CaCO3 on isopycnal surfaces in the Atlantic Ocean. Dissolution in water depths above the aragonite saturation horizon appears to be negligible. Much of dissolution occurs in depths between the aragonite and calcite saturation horizons. Dissolution rates north of 30°N are generally higher than the rates to the south, which could be partially attributable to higher production of CaCO3 in the North Atlantic than in the Atlantic side of the Southern Ocean: more CaCO3 particles rain down; more particles are subjected to dissolution. The total amount of CaCO3 that is dissolved in the Atlantic Ocean was determined by integrating estimated dissolution rates representing isopycnal surfaces throughout the entire water column and correcting for alkalinity inputs from CaCO3-rich sediments. The basin-wide dissolution rate of CaCO3 north of 30°S is approximately 0.21 Pg C yr-1, which accounts for about 64% of the net CaCO3 production for the same area. Our calculation using high quality water-column inorganic carbon data provides the first basin-scale estimate of the CaCO3 budget for the Atlantic Ocean, where the CaCO3 production rate for a given area is the highest.

Robert Daniels1, Hugh Ducklow1, George A. Jackson2, Michael R. Roman3 and Tammi Richardson2

Plankton food web structure in the NABE region, May, 1989

1 School of Marine Sciences, The College of William and Mary, Gloucester Point, VA 23062 bdaniels@vims.edu
2 Dept. of Oceanography, Texas A&M University, MS 3146, College Station, TX 77843-3146
3 Horn Point Laboratory, P.O. Box 775, Cambridge, MD 21613

We are investigating relationships between food web structure and function across different oceanic biomes using an inverse method to recover snapshots of food webs from sparse data. Specifically, we focus on how food web structure, as defined by the relative magnitude of C and N flows in a generic food web, influences particle export, nutrient regeneration, and dissolved organic carbon (DOC) cycling. Our model food web includes large and small phytoplankton, meso-and microzooplankton, bacteria, dissolved and particulate detritus, ammonium and nitrate. The majority of flows in food webs are unknown quantities that leave many questions about food web function. The inverse method, first used on plankton food webs by Vezina and Platt (1989) is a least squares approach of recovering flows from plankton food webs for which few observed data exist. The inverse method provides a solution that is consistent with the observations and with biological constraints set forth in the model. Our poster presents results of an inverse solution for a North Atlantic food web using data from the NABE (North Atlantic Bloom Experiment) study. The plankton food web was recovered for both carbon and nitrogen constituents in separate solutions. Analysis of the inverse solution shows that microzooplankton dominated the processing of carbon in the system, including grazing of phytoplankton and contributions to the DOC pool. Bacteria were very active, consuming DOC equivalent to about 60% of the net primary production. Also, the mesozooplankton export of carbon through direct fecal pellet sinking contributed to only 1% of the export.

D. DeMaster1, C. Thomas1, M. Alperin2, L. Mayer3, M. Green4, J. Aller5, R. Aller5, C. Martens2, N. Blair1, L. Benninger6 and R. Jahnke7

Studies of Organic Carbon Deposition, Recycling, and Burial on the Continental Margin Off Cape Hatteras, NC: The Ocean Margins Program

1 Dept. of Marine, Earth, and Atmospheric Sciences, No. Carolina State University, Raleigh, NC 27695
2 Dept. of Marine Sciences, University of No. Carolina, Chapel Hill, NC 27599
3 Darling Marine Center, University of Maine, Walpole, ME 04573
4 Dept. of Marine and Environmental Science, St. Josephs College, Standish, ME 04084
5 Marine Sciences Research Center, State University of New York, Stony Brook, NY 11794-5000
6 Dept. of Geology, University of No. Carolina, Chapel Hill, NC 27599
7 Skidaway Institute of Oceanography, Savannah, Georgia 31411


As part of the Ocean Margin Project, the nature and fate of organic matter reaching the continental margin seafloor off Cape Hatteras, NC has been studied. The deposition rate (or rain rate) of organic carbon to this area ranges from 3-12 moles C m-2 yr-1. Based on stable and radioactive carbon budgets, nearly all of the organic carbon is of marine origin and plankton make up from 0-30% of the organic C reaching the seafloor. Older marine organic matter that has been reworked and laterally advected to the margin comprises 60-90% of the organic carbon flux to the seafloor, whereas a refractory organic carbon source contributes 10-15%. The abundances of benthic macrofauna in continental margin sediments correlate with POC rain rate when the OMP data are combined with broader regional studies off North Carolina.

The organic carbon content of the surface sediments on the upper continental slope off Cape Hatteras ranges from 1.0 to 3.5 wt. %. These fine-grained sediments have a surface area ranging from 3 to 25 m2 g-1. Sedimentary organic carbon contents increase as grain size decreases (with water depth a secondary control). The organic loadings on surfaces of lithogenic sediment (1.5-4.5 mg OC m-2) are some of the highest measured in continental margin sediments under an oxygenated water column. Despite these high organic loadings, the mineral surfaces are essentially bare of organic coatings. High levels of dissolved organic carbon (DOC) in OMP study area pore waters appear to have little effect on the organic loading of the slope sediments. Estimates of organic carbon production by benthic fauna suggest that bacterial carbon fixation and macrofaunal carbon fixation occur at comparable rates for water depths greater than 500m. However, at depths less than 500m, where there are higher abundances of bacteria, microbial fixation rates may exceed the macrofaunal rates by a factor of 3 fold or more. The enriched D14C contents of benthic fauna from the OMP study area indicate that bomb 14C is not only making it to the seafloor, but is the predominant source of nutrition for the benthic food web. Based on 14C mass budget calculations for surface sediments, the organic matter tagged with bomb 14C comprises only a few tenths of a percent of the total organic carbon in near-interface deposits.

Bioirrigation can occur to depths as great as 200 cm (commonly 60-100 cm) in these continental margin deposits of North Carolina. The bioirrigation fluxes of dissolved inorganic carbon out of the seabed range from 2.9-7.8 moles C m-2 y-1, whereas benthic lander flux chamber experiments (with limited or no bioirrigation) yield values between 1 and 4 moles C m-2 y-1. Bromide incubation experiments on board ship corroborate the rapid transport of overlying water down into the sediment column. The penetration depths for bromide were as great as 7 cm over a period of 24 hours with the mean bromide penetration rates suggesting transport more than an order of magnitude greater than molecular diffusion.

Distributions of 234Th and chlorophyll suggest that these continental margin sediments are intensely bioturbated in the upper 7 cm over a time scale of weeks to months (Db values range from 1-200 cm2 y-1). On a 50-100 year time scale, Pu and 210Pb distributions suggest that the upper 10 to 27 cm of the sediment column are mixed more slowly (Db values ranging from 0.3 to 5.2 cm2 y-1). Burial rates of organic carbon were estimated from seabed organic carbon profiles and the distributions of Pu and 210Pb (0.1-4.7 mole C m-2 y-1 on a hundred-year time scale) as well as 14C (0.02-1.7 mole C m-2 y-1 on a thousand-year time scale). Relative to the deposition rate, the seabed preservation efficiency ranges from 3-40% (mean ~15%) with little systematic down-slope variation. In these deposits a relatively high amount of remineralized nitrogen (~68%) ultimately becomes denitrified. Burial of organic carbon in slope sediments (average rate of 0.7 moles of C m-2 y-1) accounts for only 5% of the primary production in the overlying water. Considering production and burial on the continental shelf and slope as a whole, the accumulation of organic matter on the upper slope only accounts for 0.6% of the primary production in the shelf/slope system. Despite the fact that the Ocean Margins Project site was selected to maximize the offshore transport and deposition of organic carbon, burial of organic matter in North Carolina slope sediments is not a major sink for primary production occurring in this continental margin system.

Curtis Deutsch1, John Dunne1, Jorge Sarmiento1 and Nicolas Gruber2

Diagnosing Global Oceanic Nitrogen Fixation and Denitrification

1 Atmospheric and Oceanic Sciences Program, Princeton Univ., P.O. Box CN710, Princeton, NJ 08544-0710, USA
3 Inst. Geophysics and Planetary Physics & Dept Atmos. Sci., UCLA, Los Angel es, CA 90095-1567

We present results from diagnostic simulations of the global ocean nitrogen cycle including Nitrogen Fixation and Denitrification fluxes. We begin with a simple ecosystem model in which 2 size classes of phytoplankton produce particulate and dissolved organic matter with a prognostic f-ratio. Organic matter export occurs according to a ballast based remineralization scheme after Armstrong et al. Simple parameterizations are added for nitrogen fixation and denitrification, based on local N:P stoichiometry and O2 concentration respectively. We find that total water column denitrification is too large by a factor of 2 due to extensive model anoxia. Patterns of nitrogen fixation show most new nitrate the be added in the subtropical gyres of the Pacific. Finally, nutrient uptake ratios diagnosed by the model also suggest that the Southern Ocean is a region of low N:P uptake.

Michael S. Dinniman1, John M. Klinck1 and Walker O. Smith, Jr.2

A Model Study of Circulation and Biogeochemical Processes in the Ross Sea

1 Center for Coastal Physical Oceanography, Old Dominion University, Norfolk, VA 23529 msd@ccpo.odu.edu
2 Virginia Institute of Marine Sciences, College of William and Mary, Gloucester Point, VA 23062 wos@vims.edu

Physical forcing, which includes advective circulation, vertical mixing, and vertical stratification, may be the primary factor producing the observed vertical and horizontal variability in phytoplankton distribution and primary production in the Ross Sea. Related to this, exchange of Circumpolar Deep Water (CDW) onto Antarctic Seas and continental shelves has a large influence on sea ice and biological processes. As part of the US JGOFS Synthesis and Modeling effort, we are investigating circulation and nutrient transport in the Ross Sea with an eddy permitting, regional, 3D, numerical circulation model. The present effort focuses on implementation and testing of the circulation model. Later work will consider more realistic biogeochemical processes.

We use the Rutgers/UCLA Regional Ocean Model System. Initial model fields of temperature and salinity are derived from the World Ocean Atlas (WOA98). Two different wind stress products, a monthly climatology and daily values (August 1996 to July 1997), are applied to the model. Instead of using a fully dynamic sea-ice model, ice concentrations are specified using the SSM/I climatology and this, along with the COARE bulk flux algorithm, is used to compute the model surface heat and salt fluxes. Vertical mixing in the interior and surface boundary layer is done using the K profile parameter (KPP) vertical mixing scheme (modified for the presence of ice). A radiation boundary condition is used on all the open boundaries along with adaptive nudging to monthly climatologies of tracers and volume transport. The effects of the Ross Ice Shelf are modeled by relaxing the temperature and salinity to climatological values along the edge of the shelf. Model circulation is strongly affected by bottom topography, due to weak stratification, and agrees with schematics of the general flow and long-term current measurements except near the southern boundary. There is about 2 Sv. of CDW transported onto the shelf and much of the cross shelf break transport is confined to small sections that are determined by bathymetry. The seasonal variation of the depth and temperature of the model mixed layer also match observations reasonably well.

Hugh Ducklow1, Davey Siegel2 and Bob Key3

How is SMP Doing? Perspectives on observational syntheses.

1 School of Marine Sciences, The College of William and Mary, Gloucester Point, VA 23062
2 Institute for Computational Earth System Science, University of California, Santa Barbara, Santa Barbara, CA
3 AOSP, Princeton University, Princeton, New Jersey

US JGOFS and its supporting funding agencies set an important precedent for the ocean sciences in 1996 by initiating a major program aimed at synthesizing data and understanding gained during the field phases of JGOFS. There had been smaller post-field synthesis projects funded previously but nothing on the scale of JGOFS had been attempted. The SMP was conceived as a full-fledged science program, equivalent in size, cost and scope to a major process study. In fact, the implementation process was conceived in an analogous way to a process study. Individual elements necessary for successful achievement of program goals were identified in announcements of opportunity and proposals were submitted in response. A coordinated program was composed of successful PI's. This process isn't fool-proof. Some desired projects were never proposed or funded, but some unexpected ones came along as well. As we near the final stage of SMP there have been several iterations of announcement and response, and about half the funded projects have been completed. This is a good point to ask how SMP is doing and if we have met the goals we set; and especially, to take a look at the science it has produced.

A scan through the US GOFS Blue Book (NAS, 1984) is surprising from the vantage of SMP, 16 years after the program was first imagined. Plans at the time emphasized remote sensing, primary production measurements and sediment traps. There was little explicit discussion of carbon, nor much on modeling. CO2 was nearly absent. Of course it is not surprising that the program grew and changed over 2 decades in ways few expected. Still it is instructive to look at what SMP has done to wrap up JGOFS, both as it was originally designed, and in what it became. Here, we look in particular at 3 key elements of the overall program: remote sensing, the CO2 survey, and new production/export studies, drawing mostly on completed SMP projects. A accompanying talk on modeling and prediction by Sarmiento & Dunne expands on our look at SMP.

J. P. Dunne1, R. A. Armstrong2, C. A. Deutsch1, A. Gnanadesikan1, N. Gruber3, J. L. Sarmiento1 and P. S. Swathi1

Development of a global, multi-element biogeochemical model: description, calibration and comparison with ocean color

1 Atmospheric and Oceanic Sciences Program, Princeton Univ., P.O. Box CN710, Princeton, NJ 08544-0710, USA
2 Marine Sci. Res. Ctr, SUNY, Stony Brook, NY 11794-5000
3 Inst. Geophysics and Planetary Physics & Dept Atmos. Sci., UCLA, Los Angel es, CA 90095-1567

We have developed a model to simulate ecosystem dynamics relating to regenerated production, sinking particle export and transport of dissolved organic matter in the global ocean. A key feature of this model is a representation of grazing that reproduces observed allometric relationships between large and small phytoplankton. We present an extensive compilation of data on primary production and particle export, and use this data to calibrate this relatively simple, highly parameterized model of particle export and remineralization. Production is determined by forcing nutrients toward observations. Regeneration is described as a function of temperature and community structure, competing with the sinking of detrital material through the water column. Detrital sinking is described as a function of mineral ballast. Dissolved organic matter production is described as a function of phytoplankton production and nutrient limitation and calibrated to survey data. The resulting model has been incorporated into the Princeton Ocean Biogeochemical Model to diagnose global primary production, new production, particle export and dissolved organic matter transport. Model results are compared with satellite-based primary production from ocean color to provide insights into deficiencies in the model physics and biology as well as point to potential deficiencies in the Satellite estimates.

Paul Falkowski1

General principles of marine food web structure in relation to export production

1 Institute of Marine and Coastal Science, Rutgers University, New Brunswick, NJ 08901 falko@imcs.rutgers.edu

Over the past decade, significant progress has been made in integrating measurements and models of oceanic net primary production, such that independent estimates of global carbon fixation have converged to within approximately 10% of the mean. While still far from perfect, the uncertainties in NPP are far smaller than those derived for export production. A major problem in understanding export fluxes is the application of NPZ models, that are inherently difficult to parameterize. Here I suggest that simple estimates of upper ocean mixing rates and mixed layer depths can be used to derive a latitude-dependent size structure for phytoplankton, that can, in turn, be use to constrain both estimates of export production and food web structure. Global climatologies of the particle size spectrum, in conjunction with simple models of phytoplankton functional groups, should help gain insight into herbivour community structure and food web efficiency.

R.A. Feely1, C.L. Sabine2, J.L. Bullister1 and D. Greeley1

Organic carbon remineralization rates in the Pacific Ocean

1 NOAA/Pacific Marine Environmental Laboratory; Seattle, USA
2 University of Washington/JISAO, Seattle, USA

As a part of the U.S. JGOFS synthesis and modeling project, researchers have been working to synthesize the WOCE/JGOFS/DOE/NOAA global CO2 survey data to better understand carbon cycling processes in the oceans. Working with international investigators we have compiled a Pacific Ocean data set with over 35,000 unique sample locations analyzed for at least two carbon species, oxygen, nutrient, CFC tracers, and hydrographic parameters. These data are being used determine the rates of in-situ organic carbon remineralization within the water column of the Pacific Ocean. Organic carbon remineralization rates (ranging from about 0.1 - 11 µmol kg-1 yr-1) are observed in the upper water masses from about 100 - 500 m. The rates are generally highest just below the euphotic zone and decrease with depth to values that are low and constant in the Circumpolar Deep Water. Within the North Pacific Intermediate Water (depth range: 400 - 800 m), organic carbon remineralization rates are more than 10 times higher than those observed in deepwater depths (average = 0.042 µmol kg-1 yr-1) of the Pacific Ocean.

Katja Fennel1, Mark Abbott1, Yvette Spitz1, Jim Richman1 and David Nelson1

Modeling Controls of Phytoplankton Production in the Southern Ocean -- Modern and Glacial Scenarios

1 College of Oceanic and Atmospheric Sciences Oregon State University 104 Ocean Admin. Bldg. Corvallis, Or 97331

To elucidate controls of primary and export production in the Southern Ocean we developed a one-dimensional physical/biological model. The model is applied to four stations in the southwest Pacific sector spanning the Subantarctic Zone, the Polar Front, and the Seasonal Ice Zone. The biological model component tracks the elemental cycles of nitrogen and silica. Diatoms are represented as a separate functional group. Small phytoplankton and zooplankton are tightly coupled. The one-dimensional model cannot explicitly represent horizontal fluxes of heat, freshwater and nutrients. Since these fluxes are important, we restore the temperature, salinity and nutrients in the model to available observations. We use two different approaches to include the effect of low iron availability. In modern ocean simulations, iron availability is taken into account implicitly by typical phytoplankton growth rates and a typical Si:N cell quota of diatoms. In other modern and glacial simulations, iron is included semi-explicitly, that is, iron modulates the photosynthetic efficiency and is taken up during phytoplankton growth but not tracked in the pelagic system. The model captures the essential features of the different zonal subsystems. "Top-down" control of small phytoplankton by intense grazing and "bottom-up" control of diatoms by light and silicic acid supply are the main factors for the simulated behavior. In simulations of glacial scenarios -- assuming an increase in available iron -- primary and export production increase, in particular if we assume an acclimation of the Si:N cell quota of diatoms in response to the higher iron levels.

Roger Francois

Changes in productivity, nutrient utilization and hydrology in the Southern Ocean during the last glacial maximum and its potential impact on atmospheric CO2

1 Woods Hole Oceanographic Inst., Woods Hole, MA

Following R. F. Anderson's overview of the results from the AESOPS program, I will try to synthesize the results from paleoceanographic research conducted in the southern ocean, and discuss the possible role of the biological pump of this region in controlling atmospheric CO2 on glacial to interglacial timescale. The goal will be to generate a discussion on possible new insights that the recent results from the southern ocean JGOFS program may provide to better constrain our evolving interpretation of the sedimentary record.

Marjorie A. M. Friedrichs1, Jerry Wiggert2 and Raleigh Hood3

Preliminary results from the Regional Ecosystem Modeling Testbed Project: The Arabian Sea

1 Center for Coastal Physical Oceanography, Old Dominion University, Norfolk, VA 23529
2 ESSIC, University of Maryland, College Park, MD 20742
3 University of Maryland Center for Environmental Science, Cambridge, MD 21613

The primary objective of the Regional Testbed Project is to quantitatively compare the different regional models that have been developed as part of the JGOFS SMP, in order to critically examine which ecosystem structures and model formulations are best able to simulate observed biogeochemical cycling in specific regions as well as simultaneously in multiple regions. In order to facilitate these intercomparisons we are developing a set of regional testbeds, each of which will contain one-dimensional physical forcing fields from either 3D physical model output or data, as well as biogeochemical data for either assimilation or evaluation. By running various ecosystem models using the same physical forcing, and evaluating them using the same biogeochemical data, we can objectively compare different ecosystem models and modeling approaches.

In the first year of this project, we have concentrated on formulating a prototype testbed in the Arabian Sea. In this testbed we have three distinct marine ecosystem models with varying levels of complexity, including a four-component model with diatom-like phytoplankton growth, a five-component model emphasizing the microbial loop, and an eight-component model containing multiple plankton size classes. The models are applied within a consistent one-dimensional framework at the site of the WHOI mooring (15.5°N, 61.5°E), using physical forcing fields obtained from the mooring data when possible, as well as from two three-dimensional circulation models. Chlorophyll a, nutrient and sediment trap data are assimilated using the variational adjoint method. After objectively optimizing each model in this manner, we quantitatively compare the performance of the different models to assess which model formulations best represent the fundamental underlying biogeochemical processes and capture the magnitude and variability of observed biogeochemical quantities.

Dave M. Glover1 and Maureen H. Conte1

A Coupled Epipelagic-Meso/Bathypelagic Particle Flux Model for the Bermuda Atlantic Time-series Station (BATS)/Oceanic Flux Program (OFP) Site: Phase 1, the Ecosystem Kernel

1 Dept. of Marine Chemistry and Geochemistry, Woods Hole Oceanographic Institution, Woods Hole, MA 02543 dglover@whoi.edu mconte@whoi.edu

The overarching goal of this project is to mechanistically connect euphotic zone processes with meso- and bathypelagic zone processes. It is our long term goal to accomplish this by means of a prognostic model that can be used to further our understanding of unparalleled time-series of deep-water sediment traps (21+ years) at the Oceanic Flux Program (OFP), euphotic zone measurements (10+ years) at the Bermuda Atlantic Time-series Site (BATS). In order to realize this goal we will derive a meso/bathypelagic ecosystem structure and use it to model the flux of biogeochemically active constituents (carbon, nitrogen and silica) through the water column. In this initial phase, we present the kernel of the mesopelagic ecosystem in a zero-dimensional, nitrogen only form. The equations and initial parameters are presented along with insight as to how this kernel fits into the over all scheme.

Nicolas Gruber1, Jim Orr2 and OCMIP-members, Chris Sabine3 and GLODAP members, Manuel Gloor4 and Jorge Sarmiento5

The oceanic sink for anthropogenic CO2: Combining observations with models

1 Institute of Geophysics and Planetary Physics & Department of Atmospheric Sciences, UCLA, Los Angeles, CA.
2 Laboratoire des Sciences du Climat et de l'Environnement CEA Saclay, Gif-sur-Yvette, France.
3 NOAA Pacific Marine Environmental Laboratory, Seattle, WA.
4 Max Planck Institute for Biogeochemistry, Jena, Germany.
5 AOS Program, Princeton University, Princeton, NJ.

While much progress has been made in constraining the global-scale uptake of anthropogenic CO2 by the ocean, the spatial pattern of this uptake flux is not well known. This is to a large degree due to the fact that the flux of anthropogenic CO2 across the air-sea interface cannot be separated from the often much larger air-sea flux of natural CO2. However, several independent methods have been developed over the last few years to identify the amount of total anthropogenic CO2 in ocean water as well as how this concentration changes over time (see abstract by Sabine et al.). But relating an observed accumulation of anthropogenic CO2 in the water column to an anthropogenic CO2 uptake flux at the sea surface is not straightforward, since this requires a detailed knowledge of how the large-scale circulation connects the surface ocean with the ocean's interior. I present and discuss two model-based approaches that attempt to establish this connection and thereby constrain the regional air-sea fluxes of anthropogenic CO2. The first approach is the traditional method of comparing results from forward ocean carbon model simulations with ocean reconstructions of anthropogenic CO2 and then, given reasonable agreement between the model simulated and the observed fields, arguing that the model simulated air-sea fluxes of anthropogenic CO2 are realistic. I will follow this approach using results from the recently completed 2nd phase of the Ocean Carbon-cycle Model Intercomparison Project (OCMIP). Since 13 global models provided output for anthropogenic CO2, it becomes feasible to bound the observed concentration fields with the modeled ones, thereby providing support for the argument that the real fluxes will lie within the range of the modeled ones. The second approach is a formal inversion method on the basis of Green's functions, whereby the magnitude of concentration patterns created by emitting dye tracers from a few pre-determined regions are linearly combined in such a manner that they agree with the observations optimally (in the least squares sense). Despite fundamental differences in the two methods, similar results emerge. Forward and inverse models indicate that the ocean south of 36°S takes up about 40% of the global anthropogenic CO2 uptake. The second most important region for uptake are the tropics, followed by the North Atlantic. Despite agreements in these overall pattern, substantial differences exist at more regional scales.

Nicolas Gruber1, Holger Brix1, Charles D. Keeling2 and Nicholas Bates3

Interannual to decadal variability in the carbon cycle of the subtropical gyres: A comparative study between Station 'S'/BATS and HOT

1 IGPP and Dept. of Atmospheric Sciences, UCLA, Los Angeles, CA 90095
2 Scripps Institution of Oceanography, UCSD, La Jolla, CA
3 Bermuda Biological Station for Research, Inc., Bermuda

We examine interannual to decadal variability in the ocean carbon cycle in the subtropical gyres on the basis of two long-term upper ocean time-series records. The longest record exists from near Bermuda, where sampling was initiated in late 1983 at Station 'S' and later expanded to cover the Bermuda Atlantic Time-series Station (BATS) site as well. The second record stems from the ALOHA site near Hawaii, where sampling was started with the establishment of the Hawaii Ocean Time-series (HOT) program in 1988. Both sites exhibit substantial interannual variability in all measured and computed carbon properties (dissolved inorganic carbon (DIC), total alkalinity (Alk), computed ocean surface partial pressure of CO2 (pCO2), and the 13C/12C ratio of DIC). We also find a strong anti-correlation between sea-surface temperature (SST) anomalies and DIC anomalies at both sites, which leads to a suppression of the correlation of either of these properties with pCO2. We employ a slightly modified version of the diagnostic box model of Gruber et al. [1998] to quantify the contribution of the processes controlling the carbon cycle variability at these two sites. Near Bermuda, the variability is largely driven by variations in winter mixed layer depths, which impact both the amount of DIC that gets entrained into the mixed layer and the magnitude of net community production. The variability of air-sea CO2 fluxes tends to be controlled by sea-surface temperature (SST) anomalies and accompanying wind-speed anomalies with larger CO2 uptake from the atmosphere during years of deeper than normal mixed layers. We find significant correlation of the magnitude of net community production and air-sea CO2 fluxes with the North Atlantic Oscillation (NAO), attributed to a strong influence of the NAO on convection and SST during winter. Our diagnostic analyses of the HOT data indicate a more complicated relationship between the variability in physical forcing and the response of the surface ocean carbon cycle. This is likely caused by a much weaker role played by mixed layer variations. As was the case near Bermuda, interannual variability in air-sea gas exchange is primarily controlled by variations in SST and its often associated changes in windspeed. By contrast, interannual variations in net community production are largely independent of changes in local physical forcing. Rather, we find that variability in net community production is associated with changes in horizontal advection, a process that appears to play a more important role near Hawaii than near Bermuda. This might suggest that a significant fraction of the limiting nutrient supply near Hawaii is supplied laterally rather than vertically.

Nicolas Gruber1, Hartmut Frenzel1, Patrick Marchesiello1 and J.C. McWilliams1

On the role of transport in decoupling export from new production

1 IGPP and Dept. of Atmospheric Sciences, UCLA, Los Angeles, CA 90095

One of the most important paradigms that guided the biological oceanographic community during the JGOFS period is that new production can be numerically equated with export production in steady-state situations. This permitted us to use estimates of new production as a substitute for the often more difficult measurements of export production. The assumption underlying this paradigm is that horizontal transport of organic nitrogen is neglible relative to vertical export. We investigate here the validity of this assumption using an eddy-resolving coupled physical-biological model of the central Californian coast. This system is dominated by strong coastal upwelling and generic instability of the flow regime, leading to intense formation of eddies and other meso-scale and submeso-scale features, such as jets and squirts. We find that the horizontal and vertical transports associated with such circulation structures lead to a substantial decoupling of new and export production. New production shows the expected on-offshore gradient and is primarily determined by the vertical supply of nitrate. By contrast, export production shows a complicated pattern with both negative and positive values, determined primarily by the convergence and divergence of the flow and the associated vertical transports in and out of the euphotic zone. These annual mean divergences and convergences are associated with the fact that eddies and other meso-scale features are not entirely randomly distributed, but set up an eddy-induced mean transport. Our results indicate therefore that the paradigm of numerically equal new and export production has to be used with great care, particularly in dynamic oceanographic environments.

Raleigh R. Hood1, Kevin E. Kohler2, Julian P. McCreary, Jr.3 and Sharon L. Smith4

A 4-Dimensional Validation of a Coupled Physical-Biological Model of the Arabian Sea

1 University of Maryland Center for Environmental Science, Cambridge, Maryland
2 Oceanographic Center, Nova Southeastern University, Dania, Florida
3 International Pacific Research Center, University of Hawaii, Honolulu, Hawaii
4 Rosenstiel School of Marine and Atmospheric Science, University of Miami, Miami, Florida

In this paper, we use a coupled biological/physical model to synthesize and understand observations taken during the US JGOFS Arabian Sea Process Study (ASPS). Its physical component is a variable-density, 4 and 1/2 layer model; its biological component consists of a set of advective-diffusive equations in each layer that determine nitrogen concentrations in four compartments, namely, nutrients, phytoplankton, zooplankton, and detritus. Solutions are compared to horizontal sections and time series from the ASPS data set, including observations of mixed-layer thickness, chlorophyll concentrations, inorganic nitrogen concentrations, zooplankton biomass, and particulate nitrogen export flux. Through these comparisons, we adjust model parameters to obtain a "best-fit", main run solution, identify key biological and physical processes, and identify model strengths and weaknesses. Substantial improvements in the agreement between the model and the observations are obtained by: 1) adjusting the turbulence-production coefficients in the mixed layer model to reduce bouyancy mixing; 2) changing the sinking rate and remineralization rate of detritus to to provide more rapid export and increase flux; and 3) introducing a parameterization of particle aggregation to lower phytoplankton concentrations in coastal upwelling regions. With these adjustments the model captures many key aspects of the observed physical and biogeochemical variability in the Arabian Sea. Successes include good agreement between the modeled and observed DIN concentrations and reproduction of much of the temporal and spatial variability in the mixed layer depth and phytoplankton concentrations. In addition, the model-estimated zooplankton concentrations agree with the observed mesozooplankton concentrations in offshore waters, and the model captures seasonal and spatial changes in the export flux remarkably well. Nonetheless, there are significant differences between the modeled and observed phytoplankton concentrations on virtually every cruise. In some cases these can be attributed to problems with the model's representation of the MLD variability, while in others they can be related to differences in the spatial extent of coastal upwelling, or differences in the timing of blooms between the model and the observations. Still other discrepancies can be attributed to the absence of mesoscale eddies and filaments in our relatively low resolution model. Although some problems do appear to be related to biological model, such as overestimation of phytoplankton concentrations due to underestimation of zooplankton grazing losses near the coast, we conclude that future efforts to improve the model should be focused primarily on increasing the resolution of physical model so that it can capture more of the observed mesoscale variability.

Andrew J. Irwin1 and Paul Falkowski1

Predicting Chlorophyll from Satellite and Climatological Data Products

1 Institute of Marine and Coastal Sciences, Rutgers University

An important goal in the current synthesis and modeling phase of the JGOFS project is to be able to identify the functional groups responsible for phytoplankton blooms. Some phytoplankton groups (coccolithophorids, Trichodesmium spp.) have distinct optical signatures that can be detected from space. Diatoms, the most productive of the functional groups are not uniquely identifiable on the basis of their optical characteristics. We attempt to identify diatoms using unique signatures from satellite and climatological data such as nitrate:silicate ratios. Many data sources are available including irradiance, sea surface temperature, surface wind speeds (all from satellites and with high spatial and temporal resolution), nitrate, phosphate, and silicate concentrations, upwelling velocities, mixed layer depth, aeolian dust input, and salinity (data or models, with generally poor spatial and temporal resolution). As a first step we construct a statistical model to predict chlorophyll concentration. Our best model uses 7 predictors: sea surface temperature, irradiance, irradiance / mixed layer depth, nitrate, upwelling velocity, ocean depth, and salinity. These results capture many of the spatial and temporal features of chlorophyll blooms. The model accounts for 65% of the variability in log chlorophyll and the predictions have an RMS error of 0.23 (log chlorophyll concentration). Using an increase in the derivative of the nitrate:silicate ratio as an indicator of diatoms, we estimate that 46% of chlorophyll is associated with diatoms.

Andrew J. Irwin1, Stew Sutherland2, Taro Takahashi2 and Paul Falkowski1

Using primary productivity to improve predictions of pCO2

1 Institute of Marine and Coastal Sciences, Rutgers University
2 Lamont-Doherty Earth Observatory, Columbia University

abstract The CO2 flux across the air-seawater boundary is the product of the difference in the partial pressures of CO2 and the gas exchange coefficient. Annual fluctuations in pCO2 in the ocean are driven by sea surface temperature (SST), physical mixing, and biological activity. Previous models using SST have been able to produce detailed maps of pCO2 in the surface ocean and have been used to estimate CO2 fluxes. Analysis of these models shows that the unexplained variation in pCO2 is correlated with biological activity. A statistical model using SeaWiFS chlorophyll, temperature, irradiance, and mixed layer depth is able to predict much of the residual variation in pCO2. Incorporating this biological information reduces the RMS error in pCO2 from 17 uatm to 13 umatm in the North Pacific.

M.-S. Jiang1, F. Chai1, R.C. Dugdale2, F.P. Wilkerson2, T.-H. Peng3 and R.T. Barber4

A nitrate and silicate budget in the equatorial Pacific Ocean: A coupled biological-physical model study

1 School of Marine Sciences, 5471 Libby Hall, University of Maine, Orono, ME 04469
2 Romberg Tiburon Center, San Francisco State University, 3152 Paradise Drive, Tiburon, CA 94920
3 NOAA Atlantic Oceanographic and Meteorological Laboratory, Ocean Chemistry Division, 4301 Rickenbacker Causeway, Miami, FL 33149-1026
4 NSOE Marine Laboratory, Duke University, 135 Duke Marine Lab Road, Beaufort, NC 28516

A coupled biological-physical model is developed to simulate the low silicate, high nitrate low chlorophyll (LSHNLC) condition in the equatorial Pacific Ocean. A detailed budget in the Wyrtki Box (5°N-5°S, 180°W-90°W) is carried out to understand the major sources and cycling of nitrogen and silicon in the equatorial Pacific. The modeled mean new and primary production compare well with previous observed and modeled estimates. As a major source of nutrients to the equatorial Pacific, the Equatorial Undercurrent provides slightly more nitrate than silicate to the upwelling zone, which is defined as (2.5°N-2.5°S, 180°-90°W). On the other hand, the nitrogen recycling is relatively more efficient than biogenic silicon. As a result of these combined effects, the physical supplies of silicate and nitrate into the euphotic depth in Wyrtki Box have a ratio about 0.85 (2.5 vs. 2.96 mmol/m2/day). Silicate and nitrate are taken up with a ratio of 1.17 (2.72 vs. 2.33 mmol/m2/day) within euphotic zone. The ratio of biogenic silica and nitrogen export production at the base of euphotic zone follows a 1.1:1 ratio. In the central equatorial Pacific, low silicate concentration limits diatom growth, therefore non-diatom new production accounts for most of the new production. Slightly higher silicate supply in the east maintains elevated diatom growth rate and new production associated with diatom tends to be higher. The new production associated with small phytoplankton is nearly constant along the equator. The nitrate and silicate budget calculation suggests a potential role of silicate regulation on the new production and carbon cycle in this area.

Dave Karl1

Deep secrets from the marine food web

1 School of Ocean and Earth Science and Technology, University of Hawaii, Honolulu, HI 96822

For more than a century, oceanographers have studied the interactions between the photosynthetic production of organic matter and nutrient dynamics in the sea. This research has been field-oriented and transdisciplinary, occurring at the intersections of research in microbiology, physics, analytical chemistry, cell physiology and ecology. The global data base derived from this collective effort established a sound scientific understanding of nutrient dynamics and the vital role of microorganisms, both autotrophic and heterotrophic, in the coupled organic matter production and decomposition cycles in the sea. However, novel approaches employed over the past two decades, including new designs in field experiments, repeat field observations and remote sensing capabilities, and updated methods of sample analysis, have led to a revolution in our thinking about the mechanisms and controls of nutrient dynamics in mare incognita, the hidden sea. Contemporary paradigms bear only partial resemblance to the dogma of the past, and are likely to evolve further as new data and new ideas are presented for open discussion and debate.

R.S. Lampitt1, E.E. Popova1 and I.J. Totterdell1

Downward particle flux estimates from models and measurements: The global perspective

1 Southampton Oceanography Centre, Empress Dock, Southampton, SO14 3ZH, UK R.Lampitt@soc.soton.ac.uk

Variability in upper ocean biogeochemistry determines to a large degree the temporal and spatial variations in the deep ocean downward flux of particulate material. The relevant upper ocean processes can be described by a variety of modelling approaches from simple 1D to those that are embedded in powerful general circulation models of the oceans. We describe results from a one dimensional upper ocean model at a location near to the NABE site in the Northeast Atlantic. This model is driven by meteorology to predict export flux and from that to provide a measure of downward flux at 3000m over a ten year period. The results have been found to compare very favourably with measured flux at this depth using sediment traps in terms of both the general magnitude and the characteristics of seasonal variation.

We then describe results from two similar ecosystem models that have been embedded into two general circulation models (HADOM3L and OCCAM). Both are Bryan-Cox based level models, the most important difference between them being the level of spatial resolution. The ecosystem models are of similar complexity. They have been used to derive global patterns of downward flux of organic carbon at 2000m and we compare the model outputs with each other. At 41 specific locations where long term high quality sediment trap data are available, we compare the outputs from the models with measured data. The quality of agreement between the various approaches is sometimes good but is also very variable and the reasons for this variability are discussed.

Ed Laws1

Simple solutions to complex problems

1 University of Hawaii, Oceanography Dept., Honolulu, HI 96822

Virtually all mathematical models that describe the population dynamics of aquatic food webs involve nonlinear relationships between the concentration of prey or substrate and the growth rate of the consumer. The nonlinearity is typically captured in a mathematical expression that involves the concentration of the prey or substrate and a single constant such as the half-saturation constant of the Monod equation. Because of this nonlinearity, determining the equilibrium values of the concentrations of the organisms in the food web is a superficially complex problem. I show that the nonlinearity of the differential equations can be overcome through a simple transformation that produces a set of linear equations that can be readily solved with simple matrix algebra for the equilibrium concentration values. I illustrate the technique with a simple linear food chain and with the more complex food web model of Laws et al. [Global Biogeochemical Cycles 14: 1231-1246 (2000)].

Kitack Lee1, S.D. Choi1, G.H. Park1, R. Wanninkhof2, J.L. Bullister3, R.A. Feely3, R.M. Key4, F.J. Millero5, T.-H. Peng2 and C.L. Sabine3,6

Anthropogenic CO2 in the Atlantic Ocean

1 SEE, Pohang University of Science and Technology, Pohang, Korea, email: ktl@postech.ac.kr
2 Atlantic Oceanographic and Meteorological Laboratory, NOAA, Miami
3 Pacific Marine Environmental Laboratory, NOAA, Seattle
4 AOSP, Princeton University, Princeton, New Jersey
5 RSMAS, University of Miami, Miami
6 JISAO, University of Washington, Seattle

The anthropogenic CO2 concentration in the Atlantic is separated from the large pool of dissolved inorganic carbon (CT) using a quasi-conservative quantity DC* (Gruber et al., GBC, 1996):

DC* = CTMEAS - CTEQ(S, T, ATo)| fCO2=280 uatm - RC:O2 (O2 - O2SAT) - 0.5 (ATMEAS - ATo + RNO3:O2 (O2 - O2SAT)

Where CTEQ is the CT in equilibrium with a preindustrial atmospheric CO2 concentration of 280 uatm at in situ S, T, and preformed alkalinity value, ATo. The O2 and O2SAT are in situ and saturation concentration of oxygen, respectively, S is salinity, and T is temperature (oC). The values of DC* reflect not only the excess CO2 but also the air-sea disequilibrium (DCDISEQ) at the time the water lost contact with the atmosphere. The DCDISEQ term is better estimated in this study by using the Optimum Multiparameter (OMP) analysis, which allows for more accurate determination of mixing coefficients for various water types.

This improved method was applied to new carbon dataset collected as part of World Hydrographic Program of the World Ocean Circulation Experiment (WOCE) and the Ocean-Atmosphere Carbon Exchange Study (OACES) of the National Oceanic and Atmospheric Administration (NOAA) between 1990 and 1998. In addition to the U.S. cruises, a significant number of European cruises were included in the combined dataset, which finally has data from 23 cruises. We will also present comparison of our results with those obtained by Gruber (GBC, 1998) using the TTO/NAS, TTO/TAS, and SAVE dataset.

Andrew W. Leising1, Wendy C. Gentleman1, Bruce W. Frost1, Jim Murray1 and Suzanne Strom2

Modeling the HNLC condition: some issues and constraints

1 University of Washington, School of Oceanography
2 Shannon Pt. Marine Lab, Western Washington University

One of the main issues for constructing a mechanistic-based ecosystem model is the determination of the proper transfer functions (e.g. a microzooplankton's grazing functional response) between different components or "boxes" of a model. Understanding how the choice of a particular transfer function affects a model is especially critical in oligotrophic regions or ecosystems where there are tight couplings between the microzooplankton and their phytoplankton prey, since both the uptake rates of the phytoplankton and the grazing rates of the microzooplankton may be of similar order. High-Nitrate, Low-Chlorophyll (HNLC) regions fall into this category, as phytoplankton production is co-limited by iron-limitation and heavy microzooplankton grazing pressure. Changes in the primary productivity of HNLC regions are of critical importance to atmosphere-ocean carbon flux, as HNLC regions may release large amounts of carbon to the atmosphere, rather than sequester it. Here, we show examples of some of the critical issues concerning the uptake and transfer functions within the phytoplankton-microzooplankton realm, which we believe are important for constraining more complicated models. For both phytoplankton and zooplankton, an important realization is that these organisms utilize multiple (often 2 or more) nutrient sources, which adds additional complexity to the mathematics used to describe their overall functional responses. First, we show examples of the various functions used to describe the uptake of multiple nutritional sources; these functions often sacrifice biological reality for mathematical convenience. Second, weexamined the sensitivity of one of the most common single nutrient responses for microzooplankton- the Mechalis-Menton relationship, with or without a feeding threshold - in a simple NPZ model under steady and variable physical forcing. Third, we examined a fairly complicated multiple nutrient uptake scenario involving mixotrophic algae which eat prey AND take up inorganics. Finally, we examined the historical data on phytoplankton new vs total production (as proxies for the different nutrient sources) in order to gain a better understanding of the possible mechanisms controlling these factors, and thus providing better constraints to future modeling efforts.

Ivan Lima1 and Scott Doney2

A three-dimensional, multi-nutrient, size-structured ecosystem model for the North Atlantic

1 School of Oceanography, Univ. Washington P.O. Box 355351, Seattle, WA 98195
2 WHOI, Woods Hole, MA 02543-1541

The magnitude of carbon fixation and export from the upper ocean by marine biological processes is a key variable in quantifying the flux of carbon between the ocean and the atmosphere, and our current inability to predict the ocean response to and feedbacks on anthropogenic perturbations is one of the major uncertainties for projecting future climate change. The synthesis phase of the Joint Global Ocean Flux Study has demonstrated that the combination of data from extensive field programs and remote sensing, and numerical coupled physical-biogeochemical models is a powerful tool for understanding and quantifying ocean biogeochemical processes and their potential future responses to anthropogenic perturbations. In this study, we incorporate a relatively complex ecosystem model into a three-dimensional, general ocean circulation model for the North Atlantic. The ecosystem model accounts for multi-element nutrient limitation and incorporates a more realistic, mechanistic based phytoplankton growth and photoadaptation model. Model results are compared with field data from time series stations, process oriented studies sites, and SeaWiFS imagery.

Ivan Lima1, Luanne Thompson1, Steven Emerson1 and Paul Quay1

Investigation of the physical controls of the biological pump of carbon in the subtropical North Pacific

1 School of Oceanography, Univ. Washington

The ocean plays a fundamental role in the global carbon cycle and climate system as a major sink for anthropogenic carbon from fossil fuel burning and land use change. Our current inability to predict the ocean response to and feedbacks on anthropogenic perturbations is one of the major uncertainties for projecting future climate change. The rate of mixed layer-thermocline exchange is one of the primary factors controlling the oceanic uptake of anthropogenic CO2 in the subtropical gyre. This is primarily through the solubility pump. However, the processes controlling the biological pump of carbon is especially uncertain in the large subtropical gyres. As a first step to investigate the biological pump, in this study, CFCs are incorporated as tracers into a layered isopycnal general circulation model of the North Pacific to evaluate thermocline ventilation rates. The conservation properties of the isopycnal model make it ideal for studying the physical control of tracer distributions. The physical control of CFCs will lead to better understanding of other tracers that effect the biology. The next step will be to investigate the oxygen in the model, and ultimately the nutrient supply to the euphotic zone. Preliminary results are presented and discussed focusing on the interannual variability of the thermocline ventilation and its implications for biological production and export of organic carbon.

E. Litchman1, C.A. Klausmeier2, B. van de Schootbrugge1, O. Schofield1 and P.G. Falkowski1

Applying Phytoplankton Community Models to Understanding Phytoplankton Distributions in the Paleoocean

1 Institute of Marine and Coastal Sciences, Rutgers University, New Brunswick, NJ
2 Department of Ecology and Evolutionary Biology, Princeton University, Princeton, NJ

Phytoplankton community structure has profound effects on major biogeochemical cycles in the ocean. Understanding past and present shifts in phytoplankton community composition is essential for predicting future global change. One of the biggest community shifts in the past occurred in the Mesozoic: from the ocean dominated by chlorophytes to the ocean dominated by chromophytes (diatoms, coccolithophores, dinoflagellates). What lead to the decline of green algae and rise to prominence of chromophytes?  Answering these questions may provide new insights into the modern distribution and ecological niches of the key functional groups. Here we develop a relatively simple phytoplankton community model to explore potential mechanisms of community changes in the modern ocean and the paleoocean. The model was verified with the JGOFS data from several sites. The goal of the verification procedure was to obtain qualitative agreement with the data, with an emphasis on presence or absence of certain functional/taxonomic groups, seasonal succession, nutrient concentrations and drawdown patterns. The model was then used to examine the hypotheses for the occurrence of the contrasting phytoplankton communities in the Mesozoic ocean: prasinophyte-dominated vs. dinoflagellate-dominated communities of the lower Jurassic.

Katsumi Matsumoto1, Jorge L. Sarmiento1, Andrew J. Jacobson1, Robert M. Key1, Richard D. Slater1 and Christopher S. Sabine2

Surface Bomb D14C and Surface Residence Time in Ocean Biogeochemistry Models

1 AOS Program, Princeton Univ., Princeton, NJ 08544-0710 USA
2 PMEL/NOAA, Seattle, WA 98115 USA

We compare simulations of anthropogenic CO2 and bomb radiocarbon distribution from a suite of Princeton Ocean Biogeochemical Models and models participating in Ocean Carbon Cycle Model Intercomparison Project. We show that the uptake of anthropogenic CO2 is primarily determined by vertical diffusion coefficient and secondarily by lateral eddy diffusion coefficient. Models appear to overestimate the uptake of anthropogenic tracers in the Southern Ocean, where convection is enhanced in order to simulate more correctly the deep ocean ventilation. Because radiocarbon, being a carbon isotope, has approximately a 10-year equilibration time scale, its surface concentration is determined largely by surface residence time. We formulate different methods of simulating the surface ocean age and relate it to surface bomb D14C.

D.J. McGillicuddy, Jr.1, V.K. Kosnyrev1, E.N. Sweeney1 and K.O. Buesseler1

Modeling Mesoscale Biogeochemical Processes in a Topex/Poseidon Diamond Surrounding the U.S. JGOFS Bermuda Atlantic Time-series Study

1 Woods Hole Oceanographic Institution, Woods Hole, MA 02543

An interdisciplinary modeling system has been configured in the Topex/Poseidon (T/P) "diamond" surrounding the Bermuda Atlantic Time-series (BATS) site. After extensive experimentation with the treatment of the open boundary conditions, a realistic hindcast of sea level variations in the interior of the domain has been achieved by prescribing information only along the boundaries. The time series of RMS difference between simulated and observed SLA fields for the entire altimetric record available to date shows hindcast skill that is in most cases the same order as the altimetric measurement error (3-5cm).

The T/P diamond model is being used to diagnose mesoscale biogeochemical processes in a retrospective analysis of BATS data. This activity was begun with an attempt to interpret a three-year time series record of particle flux based on thorium-234 measurements made by K.O. Buesseler. During this time period, there were three anomalously high flux events. Analysis of contemporaneous results from the T/P diamond model reveals that each of the three events took place when eddy features were present. The first two (June 1993 and August 1994) were associated with cyclonic features (negative sea level anomalies), while the last one (July 1995) was associated with with a positive sea level anomaly. Concurrent hydrographic measurements reveal the latter to be associated with a so-called "Mode water eddy," a thick bolus of 18-degree water which depresses the main thermocline and lifts the seasonal thermocline. Previous work has shown that both cyclones and Mode water eddies can inject nutrients into the euphotic zone, causing the accumulation of phytoplankton biomass in their interiors. Thus the high particulate flux events inferred from the thoriu m-234 flux measurements are consistent with these eddy-driven mechanisms.

Galen McKinley1, Mick Follows1 and John Marshall1

Interannual Variability of Air-Sea Fluxes of CO2 and O2

1 Department of Earth, Atmospheric, and Planetary Sciences, MIT 54-1517, Cambridge, MA 02139

We use an ocean general circulation model to study the mechanisms of air-sea O2 and CO2 flux variability, and consider the importance of this variability to the estimation of global CO2 sinks (Keeling et al., Nature, 358, 1996; Bender et al., Global. Biogeochem. Cycles, 10, 1996; Manning, PhD UCSD, 2001). Mean O2 and CO2 concentrations and air-sea fluxes are estimated from a multi-decadal (1980-98) model integration of the global offline MITgcm (Marshall et al., J. Geophys. Res., 120, 1997a,b; McKinley et al., Geophys. Res. Let. 27. 2000). The mean air-sea CO2 flux is consistent with the study of Takahashi et al. (Proc. CO2 in the Oceans, 1999) in all regions, and the mean O2 flux is consistent with the results of Ganachaud (ScD MIT, 1999) in all regions except the Southern Ocean.

Interannual variability in air-sea CO2 and O2 fluxes has extremes of ±0.5 PgC/y and +70/-100 Tmol/y, respectively. Globally integrated variability in O2 and CO2 air-sea fluxes is dominantly forced by the El Niño / Southern Oscillation cycle. Interannual variability of the O2 flux in the North Atlantic is also significant on the global scale. The global impact of high latitude CO2 flux variability is small.

We find the interannual variability of air-sea O2 fluxes to be large enough such that it should not be neglected in estimates of the interannual variability in land and ocean CO2 sinks based on atmospheric O2/N2 observations. In conclusion, we illustrate that estimates of CO2 sink variability from independent methods are converging toward an ocean sink variability of <1 PgC/y and a land sink variability of approximately 2 PgC/y for the 1980's and 90's.

Ben I. McNeil1, Robert M. Key1, Andrew R. Jacobson1, Louis I. Gordon2 and Jorge L. Sarmiento1

Remineralization ratios in the subsurface Indian Ocean

1 AOS Program, Princeton Univ., NJ, USA
2 College of Oceanic and Atmospheric Science, Oregon State Univ., OR, USA

The subsurface (>500m) remineralization ratios for carbon and nutrients were determined using Monte Carlo simulations of a weighted linear least squares inversion for all available data taken during the Indian Ocean WOCE program. For carbon, the anthropogenic CO2 signal was subtracted from the data while the effects of denitrification and calcium carbonate dissolution on all parameters were included in the technique. We find the remineralization ratios (P/N/O/Corg) to increase considerably with depth from 1/141/-14510/1078 in the upper ocean (500-1500m) to 1/140.5/-17515 /12510 in the deep ocean (>2000m). Although C/P and O/P increased considerably with depth suggesting preferential remineralization of phosphate in the upper water column, we did not find any significant change in N/P. The increase in C/P and C/N with depth is consistent with sediment trap results and points to possible biases in models that use constant remineralization ratios for simulations of the biological carbon pump.

Alexey V. Mishonov1, Wilford D. Gardner1 and Mary Jo Richardson1

Using the SeaWiFS data for POC assessment: which data product to use?

1 Department of Oceanography, Texas A&M University, College Station, TX, 77843

Transmissometer data were collected during six South Atlantic Ventilation Experiment (SAVE) hydrographic cruises conducted from November 1987 to March 1989 on the R/V Knorr and Melville. A total of 361 beam attenuation profiles (see Fig. 1) were made with a SeaTech transmissometer interfaced with a CTD/rosette. The regression between beam attenuation and POC for the open Atlantic Ocean waters derived from our previous research and enhanced by data from the Bermuda Atlantic Time Series (BATS) was applied in order to obtain the particulate organic carbon concentration. These data were processed and examined as vertical sections of the surface 500m. Although the data were not synoptic, data were mapped in plan view for presentation and analysis (Fig. 1). Data were integrated over the upper 30m depth for comparison with the distribution of optical data obtained from SeaWiFS. No synchronous satellite data are available for those years, but our data were compared with several satellite-derived variables from other years for comparable seasons. The highest correlation was found between POC concentration and normalized water leaving radiance at 555 nm. Other SeaWiFS-derived variables: chlorophyll concentration, diffuse attenuation coefficient at 490 nm and integral chlorophyll integrated over the upper optical depth were also processed but show less satisfactory correlation.

Mathieu Mongin1,2 and David M. Nelson2

Simulation of upper-ocean biogeochemistry in the western Sargasso Sea with a flexible-composition phytoplankton model. I. Time courses of nutrients and phytoplankton biomass in the upper 200 m

1 Institut Universitaire Européen de la Mer, Technopôle Brest-Iroise, Place Nicolas Copernic, 29280 Plouzané (France).
2 College of Oceanic and Atmospheric Sciences, Oregon State University, Corvallis, Oregon 97311, U.S.A.

We have developed a one-dimensional model of C, N and Si cycling in the upper ocean, in which biogenic material is produced by two phytoplankton groups: diatoms and nanophytoplankton. This model differs from other such models mainly in that it permits the elemental composition of the phytoplankton (C:N and Si:N ratios of diatoms and C:N ratios of nanophytoplankton) to vary freely with time and depth in response to light and nutrient availability. The growth rate of each phytoplankton group is controlled by the most limiting resource (light, N or Si) in accordance with the `cell quota' control mechanism first described by Michael Droop almost 30 years ago. Growth of each group becomes light-limited only when its C:N ratio is < 5.0 (mol/mol), N-limited only when that ratio exceeds 10.0 and diatom growth becomes Si-limited only when their Si:N ratio is < 0.6. The model thus allows phytoplankton growth and uptake of non-limiting nutrients to continue - within limits - even when light or concentrations of limiting nutrients severely limit rates of photosynthesis or nutrient uptake. Under those conditions the phytoplankton becomes deficient in C, N or Si, but continues to grow until that deficiency becomes severe enough to limit growth. Besides letting the phytoplankton respond to differential nutrient availability, the model accounts for the known differences in the light dependence of photosynthesis, nitrate uptake, ammonium uptake and silicic acid uptake.

We applied this model to the U.S. JGOFS Bermuda Atlantic Time-series Study (BATS) site for the years 1992 - 1995, a time period for which we have data on all processes considered by the model (including Si cycling and diatom productivity). Meteorological forcing was based on outputs of the European Centre Model for Weather Forecasting (ECMWF), and the physical and optical parts of the model were as developed earlier by Anderson and Pondaven (submitted). Very little `tuning' (manipulating of parameter values to obtain the best fit to the data record) was done. Instead, we chose parameter values consistent with data from phytoplankton culture studies or with field measurements at the BATS site.

The minimally tuned model reproduces the time courses of nitrate, silicic acid, chlorophyll, particulate organic carbon (POC) and biogenic silica in the upper 200 m reasonably well throughout the 1992 - 1995 time period. C:N ratios of both diatoms and nanophytoplankton are typically 9.0 - 10.0, and always significantly above the Redfield ratio of 6.6, in the upper 50 m each summer and autumn, reflecting strong N limitation. These ratios decrease to 5.0 - 6.0 the deep chlorophyll maximum (DCM) in summer and autumn, and throughout the upper water column during the spring bloom. These lower C:N ratios develop in response to greater N availability both during the bloom and in the DCM as well as lower irradiance in the DCM.

Nanophytoplankton is the dominant component of phytoplankton biomass (both as POC and as chlorophyll) at all depths and at all times. Diatom biomass reaches a maximum of ~30% of the total phytoplankton biomass in the deeper portions of the DCM in summer.

Mathieu Mongin1,2 and David M. Nelson2

Simulation of upper-ocean biogeochemistry in the western Sargasso Sea with a flexible-composition phytoplankton model. II. Primary production, nutrient uptake and nutrient regeneration

1 Institut Universitaire Européen de la Mer, Technopôle Brest-Iroise, Place Nicolas Copernic, 29280 Plouzané (France).
2 College of Oceanic and Atmospheric Sciences, Oregon State University, Corvallis, Oregon 97311, U.S.A.

The flexible-composition phytoplankton model described in Part I also provides simulations of the major biogeochemical fluxes of C, N and Si in the upper 200 m. Simulated primary productivity ranged interannually from 117.9 to 149.6 g C m-2 y-1 (mean 131.6) during the 1992 - 1995 period, somewhat lower than the 116.0 - 188.5 6 g C m-2 y-1 range (mean 154.2) estimated from 14C productivity data. Annualized f-ratios, calculated as vertically integrated annual nitrate uptake/annual (nitrate + ammonium) uptake, show little interannual change at 0.26±0.03. These f-ratios represent a composite for the upper 200 m over the year; f-ratios are lowest (0.1) in the upper 75 m in summer and autumn, significantly higher (0.2 - 0.3) within the DCM and highest (0.4 - 0.6) throughout the surface layer during the spring bloom of each year.

The model appears to require significant rates of nitrification (bacterially mediated oxidation of ammonium to nitrate) in the upper 200 m. Without this nitrification, ammonium builds up to unrealistically high concentrations (0.8 - 1.0 µM) between 100 and 170 m and nitrate is completely depleted (to < 1 nM) in the surface layer throughout the summer. The integrated annual nitrification in the upper 200 m that we estimate from model simulations ranges interannually from 222 to 310 mmol N m-2 y-1, which represents 63 - 77% of the annual uptake of nitrate by phytoplankton in the euphotic zone. Using the model outputs we have calculated an f-ratio, which is calculated as:

Nitrate uptake - Nitrification
Nitrate uptake + ammonium uptake

when all rates are integrated for the year in the upper 200 m. This f-ratio ranges interannually only from 0.048 to 0.096, implying that ~90 - 95% of the N taken up annually by phytoplankton is remineralized either to ammonium or to nitrate in the upper 200 m. If these estimates are even close to correct, our simulations tend to support the conclusion of Lipschultz (2001) that nitrate is not a purely `new' nutrient in the upper 200 of in the western Sargasso Sea.

The simulations imply that production of biogenic silica by diatoms at the BATS site ranges interannually from 162 to 247 mmol Si m-2 y-1 (mean 204), somewhat lower than the data-based estimate of ~240 mmol Si m-2 y-1 made by Nelson and Brzezinski (1997). These silica fluxes imply that diatoms are responsible for ~25 - 30% of the annual primary production at the BATS site. Perhaps surprisingly, the diatoms' contribution to annual new production is considerably lower (14 - 19%) as a consequence of their assumed higher KS for nitrate uptake.

J. Keith Moore1,

Modeling Southern Ocean Carbon Fluxes

1 University of California, Irvine, Dept. of Earth System Science, Irvine, CA. jkmoore@ucar.edu

A decade ago the Southern Ocean was one of the most under-sampled, under-studied regions of the world ocean. Since then there have been numerous in situ studies in this region that have provided modelers with the critical data needed to constrain/evaluate model output. This talk will begin with a satellite overview of the Southern Ocean focusing on regional variability and the role of Southern Ocean fronts. Model estimates of atmospheric iron deposition to the Southern Ocean within mineral dust for modern times and at the Last Glacial Maximum (LGM) will be presented. Sensitivity to variations in dust deposition using a global mixed surface layer ecosystem model (Moore et al., 2002) will be examined. Preliminary results from a global 3D ecosystem model will be used to look at how the factors limiting phytoplankton growth in the Southern Ocean vary over seasonal timescales. Lastly, differences in dust deposition and sea ice cover at the LGM will be discussed in terms of how these may have influenced carbon cycling in the Southern Ocean.

J. Keith Moore1, Scott Doney2 and Keith Lindsay3

Phytoplankton Functional Groups and Oceanic Carbon Cycling

1 University of California, Irvine, Dept. of Earth System Science, Irvine, CA. jkmoore@ucar.edu
2 Woods Hole Oceanographic Institution, Woods Hole, MA. sdoney@whoi.edu
3 National Center for Atmospheric Research, Boulder, CO.

A modified version of the Moore et al. (2002) marine ecosystem model that includes several key functional groups of phytoplankton and allows for multiple potentially limiting nutrients has been incorporated into the ocean component of the NCAR Community Climate System Model. The ecosystem model is coupled with a full biogeochemical module that includes carbonate system dynamics and air-sea gas exchange of oxygen and carbon dioxide. Phytoplankton functional groups represented include diatoms, nitrogen-fixing diazotrophs, coccolithophores, and picoplankton. Phytoplankton growth rates are a function of available light, nitrogen, phosphorus, iron and (for the diatoms) silicon. The inclusion of an explicit iron cycle, including the atmospheric source from dust deposition, allows the model to capture the observed High Nutrient, Low Chlorophyll conditions in the subarctic and equatorial Pacific, and in the Southern Ocean. Model results are compared with global in situ nutrient measurements and satellite-based estimates of surface chlorophyll concentrations and primary production. Controls on phytoplankton growth rates at the global scale are examined.

Ragu Murtugudde1 and Jim Christian1

Decadal variability in the tropical ecosystems

1 ESSIC, Univ of MD, College Park, MD 20742

While the tropical Pacific is dominated by ENSO, the Atlantic has zonal and meridional modes of variability on interannual time-scales. An interdecadal regime shift has been noted in the Pacific and studied widely for its signatures in the marin ecosystem from primary producers to fisheries. It has been shown recently that a simultaneous regime shift also occurs in the Atlantic and Indian Oceans in terms of a thermocline/nitracline shift. The effects of this shift on the marine ecosystems and the characteristics of this decadal variability are studied in a coupled biological-physical model. Most of the variability at these time-scales is associated with the large phytoplankton suggesting consequences for the oceanic carbon cycle. The tropical Atlantic appears to have undergone a change in its primary model of interannual varaibility from a predominantly meridional to zonal or Atlantic Niño model after 1976. The stronger and more frequent El Niños in the Pacific are obviously manifest in the ecosystem response in terms of reduced primary production in the east and enhanced entrainment in the western Pacific warm pool. The details of the model simulations of the circulation and ecosystem are presented.

Payal Parekh1*, Mick Follows1 and Ed Boyle1

Controls on deep water iron distribution

1 M.I.T., Department of Earth, Atmospheric, and Planetary Sciences, *MIT/WHOI Joint Program

Waters upwelling to the surface ocean are deficient in iron relative to other nutrients in terms of the requirements for biological productivity. The extent of this deficit, and whether it may be compensated locally by aeolian deposition, depends upon the balance of ocean transport and biogeochemical mechanisms. We examine mechanisms which may control the global, deep ocean iron distribution in the context of a simplified ocean transport and biogeochemistry model. The six box global ocean model includes coupled phosphorous and iron cycles with prescribed atmospheric deposition of iron and ocean circulation. Export production is parameterized simply, limited by phosphate and iron availability in the surface ocean.

We compare several parameterizations of the deep water, geochemical processes affecting iron based on scavenging and backscavenging, analagous to the inferred properties of thorium, and complexation, the binding of iron to organic ligands. If complexation is not represented, provided that a ratio of scavenging to backscavenging rates of ~.01 is chosen the scavenging-backscavenging model can reproduce the observed, basin to basin, deep water gradients of iron. The introduction of a strong ligand into the model changes the balances, reducing the significance of backscavenging. The observed deep water gradients can still be reproduced. In addition, this model predicts that over 95% of the iron is organically bound and also leads to the presence of excess, uncomplexed ligand, in agreement with observations, but not captured in previous global models. Sensitivity tests indicate that deep water Fe gradients are sensitive to the scavenging rate and total ligand concentration. The observed iron distribution can be reproduced with a broad range of total ligand concentrations in combination with the appropriate scavenging rate.

These simplified models reinforce the concept that the deep water iron distribution is controlled by a complex balance of ocean transports, complexation, scavenging onto particles, and regionally varying aeolian dust deposition. The idealized model provides a testbed for examining basic parameterizations and in which to perform sensitivity studies. The study also suggests a strategy for modeling the iron cycle, and its effect on regional productivity, in global biogeochemical models.

Tsung-Hung Peng1 and Yuan-Hui Li2

Penetration of Anthropogenic CO2 in the Oceans Based on Analysis of Recent WOCE/JGOFS/OACES Carbon Data Using the Remineralization Ratios Obtained by the New Three-End-Member Mixing Model

1 NOAA/AOML, Ocean Chemistry Division, Miami, FL 33149
2 Department of Oceanography, University of Hawaii, Honolulu, HI 96822

In a recent report (Li and Peng, 2002), a new three-end-member mixing model is used to obtain remineralization ratios of organic matter in the water column. Remineralization ratios (P/N/Corg/-O2) of organic matter in the deep water column change systematically from the northern Atlantic to the Southern Oceans, then to the equatorial Indian and the northern Pacific oceans, more or less along the global ocean circulation route of deep water. Average remineralization ratios of organic matter for the northern Atlantic Ocean are P/N/Corg/-O2 = 1/(16±1)/(73±8)/(137±7), and for the Southern Oceans P/N/Corg/-O2 = 1/(15±1)/(80±3)/(133±5). Those values are similar to the traditional Redfield ratios of P/N/Corg/-O2 = 1/16/106/138 for marine plankton, except for the low Corg/P ratio. Average remineralization ratios for the equatorial Indian Ocean are P/N/Corg/-O2 = 1/(10±1)/(94±5)/(130±7), and for the northern Pacific Ocean P/N/Corg/-O2 = 1/(13±1)/(124±11)/(162±11). The apparent low N/P ratio for both ocean basins suggests that organic nitrogen was converted partly into gaseous N2O and N2 by bacteria through nitrification/denitrification processes in a low-oxygen or reducing micro-environment of organic matter throughout the oxygenated water column. The actual N/P ratio of remineralized organic matter is probably around 15±1. The O2/Corg ratio of remineralized organic matter also decreases systematically along the global ocean circulation route of deep water, indicating changes in relative proportions of biomoledules such as lipids, proteins, nucleic acids, and carbohydrates.

In contrast, uniform remineralization ratios are assumed for almost all the current calculations using C* method for the anthropogenic CO2 inventory in the three major oceans (for example, Gruber et al., 1996; Gruber, 1998; Sabine et al., 1999 and 2002; and Feely et al., 2001). The question that needs to be asked is: how much difference in anthropogenic CO2 penetration into the ocean could be made if the new results of variable ratios are used as compared with those of traditional methods where the uniform remineralization ratios have been used. In this study, a new method derived from our mixing model, which is different from C* method, is developed for estimating the penetration of anthropogenic CO2 into the ocean, specifically using the current variable remineralization ratios. This new method will be described in details, and comparison of current results based on HOT station north of Hawaii will be made between using uniform remineralization ratios and variable remineralization ratios.


Feely, R. A., C. L. Sabine, T. Takahashi, and R. Wanninkhof, Uptake and storage of carbon dioxide in the ocean, Oceanography 14,18-32, 2001.

Gruber, N., j. L. Sarmiento, and T. F. Stocker, An improved method for detecting anthropogenic CO2 in the oceans, Global Biogeochem. Cycles 10, 809-837, 1996.

Gruber, Anthropogenic CO2 in the Atlantic Ocean, Global Biogeochemical Cycles, 12, 165-191, 1998.

Li, Y.-H. and T.-H. Peng, Latitudinal change of remineralization ratios in the oceans and its implication for nutrient cycles. Global Biogeochem. Cycles, (submitted) 2002.

Sabine, C.L., R.M. Key, K.M. Johnson, F.J. Millero, A. Poisson, J.L. Sarmiento, D.W.R. Wallace, and C.D. Winn, Anthropogenic CO2 inventory of the Indian Ocean, Global Biogeochem. Cycles, 13, 179-198, 1999.

Sabine, C. L., R. A. Feely, R. M. Key, J. L. Bullister, F. J. Millero, K. Lee, T.-H. Peng, B. Tilbrook, T. Ono, and C. S. Wang, Distribution of anthropogenic CO2 in the Pacific Ocean, Global Biogeochem. Cycles, (in press) 2002.

B. B. Prézelin1, E. E. Hofmann2 and J. M. Klinck2

Physical Forcing of Phytoplankton Community Structure in Continental Shelf Waters of the western Antarctic Peninsula

1 Mar. Sci. Inst. and Dept Ecology, Evolution and Marine Biology, Univ, California, Santa Barbara, CA 93106
2 Center for Coastal Phys. Oceanogr., Old Dominion Univ. Norfolk, Virginia 23529

A previous study of the Western Antarctic Peninsula (WAP) continental shelf that was based upon a multidisciplinary data set collected during austral summer of January 1993 identified a mechanism previously unrecognized that sets up a physical and chemical structure that supports enhanced biological production (Prézelin et al. 2000). This biological production occurs when the southern boundary of the Antarctic Circumpolar Current (ACC) flows along the shelf edge, which produces upwelling and intrusions of nutrient-rich Upper Circumpolar Deep Water (UCDW) onto the shelf, thereby allowing site-specific diatom-dominated phytoplankton communities to develop. The enhanced biological production potentially affects all components of the marine food web in this region. In this analysis, we extend the area and seasons studied through similar analyses of multidisciplinary data sets collected on four additional cruises that cover all seasons. We find that this newly recognized forcing is active in other regions of the WAP shelf where similar conditions are found, is episodic, and is forced by non-seasonal physical processes. The meander frequency of the ACC has consequences for the timing and location of UCDW intrusions. When multiple intrusions are observed, each event may be in a different stage. Further, the occurrence of an event in one area does not necessarily imply that similar events are ongoing in other areas along the shelf. While these UCDW upwelling events originate along the outer shelf, they have a signature that extends into the inner shelf region because of the deep topography with allows the inner shelf to be connected to the outer shelf. The frontal boundary between the intruded water and the shelf water is variable in location because of the episodic nature of the onshelf intrusions, being moved further inshore when one of these events is occurring. The frontal boundaries are characterized by distinct phytoplankton communities whose distribution along the circulation structure is identifiable by the unique presence of a chemotaxonomic marker (Chlorophyll b) in the near surface waters. These observations show clearly that the phytoplankton community structure on the WAP shelf is determined by physical forcing. Moreover, variability in this physical forcing, such as may occur via climate change, can potentially affect the overall biological production of the WAP continental shelf system.

Donald G. Redalje1, Steven E. Lohrenz1, Gustav-A. Paffenhöfer2, Peter G. Verity2 and Charles N. Flagg3

Budgets of Biogenic Elements in the NW Atlantic Ocean Margin: A Synthesis and Modeling Project - Major Findings of the Water Column Process Group

1 The Univ. of Southern Mississippi, Dept Marine Science, Stennis Space Center, MS 39529
2 Skidaway Inst. of Oceanography, Savannah, GA 31411
3 Brookhaven National Lab, Upton, NY 11973

The central objectives of our program were (1) to quantify the processes and mechanisms that affect the cycling, flux, and storage of carbon and other biogenic elements at the land/ocean interface; (2) to define the ocean-margin sources and sinks in global biogeochemical cycles; and (3) to determine whether ocean margins, including continental shelves, are quantitatively significant in removing carbon dioxide from the atmosphere and isolating it via burial in sediments or by export to the interior ocean, or elsewhere. The objectives of the research presented here were to examine the biomass of organisms that participate in the cycling of CO2 and the rates of production and grazing that will enable us to address the overall program objectives presented above.

The field study involved an experimental approach that included sampling over a broad range of spatial and temporal scales. Sampling approaches included (1) "shelf-wide" survey cruises (2 in late Winter/Spring and 1 in Summer) from Nantucket to Cape Hatteras; (2) long-term deployment (most of 1996) of an array (26 moorings) of instrumented moorings between Chesapeake Bay and Cape Hatteras; and (3) late Winter/Spring and Summer detailed process cruises that included using drogues to allow for the time-series sampling of single water masses that flowed through the instrumented moorings (the Lagrangian experiment conducted in March and July 1996) as well as cross-shelf surveys conducted along the axes of the mooring arrays during each process cruise.

Using information derived from the process cruises, areal estimates of vertically integrated primary production were generated using a wavelength resolved photosynthesis-irradiance model to generate production estimates for the entire duration of the program. In addition, vertically integrated primary production was measured and detailed grazing experiments were conducted during each of the Lagrangian experiments. Biomass estimates, ranging from pico- and nanophytoplankton to mesozooplankton, and rates of transfer from dissolved CO2 through the various groups of organisms have been determined. During the Lagrangian experiment we also determined growth rates for both the phytoplankton and the microzooplankton present in the water mass. The biomass and rate processes of the phytoplankton and microzooplankton community were sufficient to support biomass abundances of mesozooplankton that rank as high as any ever reported for continental shelves and slopes.

Major findings of the Water Column process Group include:

Rates of primary production in the study area varied from 0.5 - 0.6 gC m-2 d-1 during March to 0.4 - 2.1 gC m-2 d-1 in July, with > 50% of this production attributed to the < 8 mm size-fraction. Microzooplankton consumed 65% of the chlorophyll a production for the whole community and 81 % of the < 8 mm size-fraction. Due to the observed high rates of production and consumption of the < 8 mm size-fraction, it is clear that in our study region the smaller autotrophs contributed significantly to production and that this production was actively incorporated into the microbial food web. Rates of phytoplankton and growth as well as microzooplankton herbivory were a function of temperature. Thus we observed low phytoplankton growth rates in March, when water temperatures averaged about 7°C, even though primary production was 0.5 - 0.6 gC m-2 d-1; biomass levels were sufficient to support observed levels of production and growth.

Mary Jo Richardson1, Wilford D. Gardner1 and Alexey V. Mishonov1

Synthesis of the POC field in the Pacific based on historic WOCE transmissometer data

1 Department of Oceanography, Texas A&M University, College Station, Texas, 77843

Transmissometer data collected during several WOCE expeditions in the Pacific from 1991 to 1994 have been analyzed and processed to investigate the spatial distribution of POC and the impact of El-Niño on that distribution. The relationship between in-situ POC concentration and beam attenuation was determined from a selected set of Hawaii Time-Series data. This relationship (see Fig. 1) was used to calculate the basin-wide POC distribution in the upper 500 m. The El-Niño (1991-1994) events can be seen clearly in the POC sections, characterized by low values of POC in the upper ocean layer in the equatorial region.

Tammi L. Richardson1, George A. Jackson1, Hugh W. Ducklow2 and Michael R. Roman3

Inverse Analysis of Planktonic Food Webs of the Equatorial Pacific and Arabian Sea

1 Dept. of Oceanography, Texas A&M University, MS 3146, College Station, TX 77843-3146
2 Virginia Institute of Marine Science, P.O. Box 1346, Gloucester Point, Virginia 23062-1346
3 Horn Point Laboratory, P.O. Box 775, Cambridge, MD 21613

This SMP-funded research focuses on ecosystem structure, biogeochemical fluxes, and vulnerability to climate change perturbations in the ocean. Our overall research goal is to synthesize food web data from JGOFS studies to understand the mechanistic controls of particle consumption and export, nutrient regeneration and DOC production/export in oceanic systems. We will use this information to examine stability properties and thus assess the vulnerability of different food web structures to ocean warming.

Our poster presents initial results of this study. We have consolidated field measurements from the Equatorial Pacific and Arabian Sea into descriptions of size- and function-based food webs and have used inverse analyses to infer unknown (i.e. unmeasured) flows of material within the webs. For the Equatorial Pacific, food webs for two Time-Series cruises (spring and fall 1992) are presented. For the Arabian Sea, we present results from four analyses, including an offshore and near-shore station for two seasons, the spring Inter-Monsoon and Southwest Monsoon of 1995.

Paul Robbins1

Estimates of Oceanic Anthropogenic Carbon based on Chloroflourocarbon Inventories: Applying generalized aged distributions to ocean ventilation

1 Physical Oceanography Research Div., Scripps Inst. Oceanography, Univ. of California, San Diego, 9500 Gilman Drive, Mailcode 0230, La Jolla, CA 92093-0230

A new technique is introduced to estimate total anthropogenic carbon inventories in the ocean based on CFC inventories calculated from hydrographic observations. The method offers two significant advances. First, the CFC observations are interpreted in the framework of Age Distributions rather then being treated as measures of a single discrete transit time, thus relaxing previous assumptions which neglect internal mixing within the ocean. Second, bias uncertainties due to both assumptions in the method and mapping errors can be explicitly calculated.

Christopher L. Sabine1, R.A. Feely2, R.M. Key3, R. Sonnerup1, B. McNeil3, K. Lee4, and N. Gruber5

Global Estimates of Anthropogenic CO2: Current Methods and Results

1 JISAO, University of Washington, Seattle, WA, USA
2 NOAA/PMEL, Seattle, WA, USA
3 AOS Program, Princeton Univ., Princeton, NJ, USA
4 School of Environ. Science and Engineering, Pohang, South Korea
5 IGPP & Dept. of Atm. Sciences, UCLA, Los Angeles, CA, USA

Significant advances have been made in isolating the anthropogenic component of ocean TCO2 concentrations during the JGOFS era. Several techniques have been proposed and used in various regions. Some of these approaches estimate the total anthropogenic concentration, like the Gruber et al. DeltaC* technique or the Goyet et al. Mix approach; others estimate the anthropogenic increase between two cruises, like the Wallace excess CO2 technique or the McNeil et al. CFC approach. Some of the approaches estimate anthropogenic CO2 concentrations based primarily on inorganic carbon measurements while others primarily use changes in proxy tracer concentrations or age spectrums. Each of the techniques involves a finite set of assumptions that, in many cases, are similar between techniques. The different techniques have differing strengths and weaknesses depending on these assumptions and the data quality. Most of the current analyses of the WOCE/JGOFS global survey data have used the DeltaC* technique. A summary of the global anthropogenic CO2 estimates based on this approach will be presented. We will also briefly review some of the other more popular approaches for estimating anthropogenic CO2 in the oceans, discussing the strengths and weaknesses of each. Where possible, these techniques will be directly compared with the DeltaC* approach, with historical approaches for estimating anthropogenic CO2, and with the recent OCMIP II results.

C.L. Sabine1, R.A. Feely2, R.M. Key3, and D. Greeley2

Geochemical Evidence of Carbonate Dissolution in the Pacific and Indian Oceans

1 Joint Institute for the Study of the Atmosphere and Ocean, University of Washington, Seattle, Washington, USA
2 Pacific Marine Environmental Laboratory, NOAA, Seattle, Washington, USA
3 Atmospheric and Oceanic Sciences Program, Princeton University, Princeton, New Jersey, USA

Over the past several years we have been working to synthesize the WOCE/JGOFS/OACES global CO2 survey data to better understand carbon cycling processes in the oceans. The Pacific and Indian Ocean data sets have over 60,000 sample locations with at least two carbon species, oxygen, nutrients, chlorofluorocarbons, and hydrographic parameters. Extensive quality assessments of these data suggest that the dissolved inorganic carbon (DIC) and total alkalinity (TA) data are accurate to approximately ±3 and ±5 µmol kg-1, respectively. These data are used to examine the geochemical signatures of calcium carbonate dissolution in the Pacific and Indian oceans.

For a given isopycnal surface, the changes due to CaCO3 dissolution can be evaluated using what we term TA*:

TA* = 0.5 (NTA - NTAO) + 0.63 (0.0941 × AOU)
where NTA = (TA × 35)/S, NTAO is a preformed alkalinity based on a linear regression of surface TA values against salinity and potential temperature, and AOU is the apparent oxygen utilization. TA* concentrations start to increase in relatively shallow waters near or slightly above the aragonite saturation horizon. Below this horizon, TA* concentrations increase rapidly from about 10 to 40 µmol kg-1. The highest concentrations are observed at intermediate depths at the northern end of both oceans. The exact mechanisms that promote this dissolution still need further investigation, but the water chemistry analyses are consistent with a rapid dissolution of carbonate particles just below the aragonite saturation horizon. The measurable upward migration of the aragonite and calcite saturation horizons since the pre-industrial period suggest that the long-term impacts of CaCO3 dissolution on the oceans ability to neutralize anthropogenic CO2 need to be considered in future biogeochemical models.

Baris Salihoglu1 and Eileen E. Hofmann1

Development of a Complex 1D Ecosystem Model for the Equatorial Pacific.

1 Center for Coastal Physical Oceanography, Old Dominion University, Norfolk VA, USA baris@ccpo.odu.edu

A complex ecosystem model is developed to investigate the biological and physical interactions in the equatorial Pacific Ocean at 140°W. In particular the model is used to investigate the mechanisms that lead to seasonal shifts in phytoplankton species composition. The model is forced by seasonal changes in spectral light, temperature, and water column mixing. Autotrophic growth is represented by four algal groups of phytoplankton. The groups have light and nutrient utilization characteristics that reflect those of Prochlorococcus, Synechococcus and Chromophycota species. The model results indicate shifts in the species composition of phytoplankton appear in the simulated distributions as depth and season changes. The extreme light-inhibited Prochlorococcus becomes the most abundant group in deeper waters (as light intensity decreases). During La Niña the largest algal group, Chromophycota, becomes the most abundant, whereas, it was the least abundant group during El Niño.

J.L. Sarmiento1 and J. Dunne1

How is SMP Doing? Perspectives on modeling and prediction

1 Atmospheric and Oceanic Sciences Program, Princeton Univ., P.O. Box CN710, Princeton, NJ 08544-0710, USA

abstract pending The original SMP Implementation Plan of February 1997 identified the following goal and program elements related to modeling and prediction:

Goal: To improve our ability to predict the role of oceanic processes and feedbacks in determining the future partitioning of carbon between the ocean and atmosphere, and to evaluate uncertainties and identify gaps in our knowledge.

Program Elements: Mechanistic Controls: principal processes that control within-ocean and ocean-atmosphere partitioning of carbon and related chemicals, with a view towards developing regional and global syntheses and models. Extrapolation and Prediction: development and application of methods that will allow knowledge gained on small spatial and temporal scales to be scaled to seasonal, annual, and interannual time scales and to regional and global spatial scales, and development and application of methods that will improve our ability to predict the role of oceanic processes and feedbacks in determining the partitioning of carbon between the ocean and atmosphere.

A more specific statement of the goal is given on the SMP web site (S. Doney, J. Kleypas, and J. Sarmiento): "The concept of encapsulating the findings of JGOFS in the form of numerical models has been a stated goal from the inception of the overall program. One proposed element as part of the "grand synthesis" of SMP would be a "core biogeochemical model" implemented globally in a 3-D circulation model. Rather than promoting a single model that could address all scientific questions (such a model is a fictional chimera at best), the SMP could develop a base-line global biogeochemical model of manageable complexity and computational demands that incorporates the emerging new understanding from JGOFS and the SMP."

We provide an overview of progress towards meeting the goal, starting with the development of mechanistic models. Some of the particular areas of progress in representation of mechanisms include: 1) improved representation in models of the role of small and large phytoplankton in determining biogeochemical cycling, 2) improvement in our understanding of and ability to model the role of iron, though much still remains, 3) models of the role of ballast in determining export from the surface ocean and through the ocean interior, 4) increasing understanding and incorporation into models of the role of functional groups in determining production, export and variability in elemental stoichimetry, and 5) identification of the importance of temporal variability and mesoscale physical processes in nutrient supply. The second point that we discuss is our present understanding of the response of the global carbon system to perturbations in the ocean carbon cycle, including: 1) the nature of the differences between box models and GCMs, 2) quantification of the potential for iron fertilization to draw down atmospheric CO2 and 3) for deep injection of CO2 to limit future growth of atmospheric CO2, and 4) response of the carbon cycle and biology to global warming.

Reiner Schlitzer1

Estimating Ocean Carbon Fluxes using Inverse Methods

1 Alfred Wegener Institute for Polar and Marine Research, Columbusstrasse, D 27568 Bremerhaven, Germany; rschlitzer@awi-bremerhaven.de

In recent years large, globally integrated datasets related to marine biogeochemical cycles have been released to the public and are now in wide use. Large observational programmes such as WOCE and JGOFS, which are completed in the near future will provide more data resources of unprecedented quality and extent. Availability and easy access to extensive, global databases will likely foster the use of inverse methods that attempt to estimate physical transports and biogeochemical fluxes and reaction rates from the available data. However, it is common understanding that even the large datasets of the future will be inadequate to fully determine biogeochemical fluxes on a global scale. Heterogeneity of the different data types and uneven and incomplete spatial and temporal data coverage are additional challenges demanding rather complex mathematical methods and simplified models.

As an example, an inverse model approach that uses water-column distributions of hydrographic parameters, oxygen, dissolved nutrients, carbon and transient as well as steady-state tracers to determine the underlying production, sinking and remineralization of particulate and dissolved organic carbon (POC and DOC) is described in detail. A comparison with satellite-derived productivity estimates shows large discrepancies in the Southern Ocean south of 50°S, where model export of POC are about a factor of 2 larger than satellite estimates. Possible explanations for this discrepancy are a potentially poor calibration of satellite sensors and productivity algorithms in this region and the difficulties to detect and/or parameterize frequently observed sub-surface chlorophyll patches.

Walker O. Smith, Jr.1, Michael S. Dinniman2, John M. Klinck2 and Eileen Hofmann2

Biogeochemical Climatologies in the Ross Sea, Antarctica: Modeling the Temporal Patterns of Primary Production

1 Virginia Institute of Marine Science, College of William and Mary, Gloucester Pt., VA 23062
2 Center for Coastal Physical Oceanography, Old Dominion University, Norfolk, VA 23529

The temporal pattern of nutrient (nitrate and silicic acid) and chlorophyll distributions in the Ross Sea is formulated by two independent ways. The first procedure compiles all available data from known cruises from 1970 to the present and generates a three-dimensional grid for the months from November through February using the iterative difference-correction scheme. The second method uses a three-dimensional circulation model that includes the effects of the Ross Sea gyre off the continental shelf, and investigates the effects of currents and phytoplankton uptake on nutrient distributions and phytoplankton standing stocks. The two approaches produced similar results, although the circulation model produced distributions that were more variable in space due to its finer resolution. The nutrient distributions were characterized by elevated concentrations in early spring and gradual reductions to ca. 15 and 40 M (nitrate and silicic acid, respectively) in summer. Nutrient depletion did not occur despite the favorable growth conditions in summer, suggesting that an alternative limitation occurs. Modeled chlorophyll concentrations reached ca. 5 g L-1 in December and declined thereafter, similar to the observed in situ seasonal declines. Seasonal primary production calculated from the nitrate deficits and the circulation model suggested that production was ca. 120 g C m-2, similar to other estimates using independent methods. Both the nutrient/pigment climatologies and circulation model results confirm that the Ross Sea continental shelf is among the most productive regimes of the entire Southern Ocean.

Yvette H. Spitz1, Mark M. Abbott1 and James G. Richman1

Analysis of the three-dimensional coupled circulation-ecosystem model results for the North Pacific Ocean

1 Oregon State Univ., College of Oceanic and Atmospheric Sciences, Corvallis, OR 97331

As part of the NASA funded Carbon Cycle program, we proposed to quantify the impacts of changing community structure and to simulate the processes relevant to the biologically-driven carbon export in the context of a basin-scale model of the North Pacific. In that context, we also proposed to assimilate SeaWiFS chlorophyll-a measurements and quantify the errors of model-derived estimates of carbon flux. As a first step to achieve our goals, we must quantify the behavior of the circulation as well of the ecosystem model. In addition, we need to have thorough understanding of the spatial and temporal variability of the chlorophyll-a field. Using the 8 day composite 9 km x 9 km ocean color images from 1998 to 2001, we analyzed the yearly and seasonal mean and standard deviation patterns. The ecosystem model is coupled to a general circulation model (OGCM) of the North Pacific. The OGCM is based upon the Parallel Ocean Program (POP) model of the Los Alamos National Laboratory. The circulation model reproduces the large-scale behavior of the oceanic mixed layer very well. The SST at two time series locations, the Hawaiian Ocean Time Series (HOT) and the TOA buoy on the equator at 140°W, and the corresponding model estimate agree remarkably well. The model and observations are correlated at 0.91 at HOT and 0.85 at the TAO mooring. The model appears to be biased slightly warmer than the observations. Basin-wide, the variability in the SST is dominated by the seasonal cycle. The behavior of the upper thermocline leads, however, to surprising result. The dominance of the seasonal cycle in the upper 100m is evident, even at the tropical HOT site. Below the mixed layer, a substantial warming trend is observed in the 1980s with the 20°C isotherm deepening from 117m in the Levitus initial condition to approximately 160m in 1990. In the 1990s, the deepening trend disappears with the 20°C isotherm fluctuating around 170m. The surface chlorophyll-a patterns observed in the seasonal mean (April-June 1988-1991) of the model results are in good agreement with the seasonal mean (April-June 1998-2001) from SeaWiFS. Comparison between a one-dimensional and a three-dimensional ecosystem simulation shows that horizontal advection and diffusion play an important role in the interannual variability observed at HOT. While the 1D simulation leads to the right order of magnitude and location for the deep maximum chlorophyll-a, this simulation is unable to reproduce the interannual variability observed in the data set between 1989 and 1991. On the contrary, the 3D simulation displays a similar interannual variability to the observations. The deep Chl-a is elevated in 1989 then decreases in 1990-1991 and increases in 1993-1994. The primary productivity displays similar temporal evolution in the 3D simulation and in the observations while the increase of surface primary productivity in 1991 is not apparent in the 1D simulation. The 1D simulation reproduces, however, the decadal variability with some accuracy. For example, the 1D model results display an increase of the DOC pool from 1988 to 1997 and a slight decrease after 1997, as seen in the observations.

Colm Sweeney1, Kevin Arrigo2 and Gert van Dijken2

Prediction of pCO2 in the Ross Sea, Antarctica Using Ocean Color Data

1 Lamont-Doherty Earth Observatory, RT 9W, Palisades, NY
2 Stanford Univ., Mitchell Building, Stanford, CA 94305

A time series of the surface partial pressure of CO2 (pCO2) made from underway samples on the R/V N. B. Palmer from the austral spring of 1996 through mid-summer of 1999 in the Ross Sea, Antarctica are compared with the predicted pCO2 derived from climatological mixed layer depths and estimated primary productivity using observations of ocean color (OCTS and SeaWiFS). Both insitu measurements and predicted values of pCO2 indicate a large net biological drawdown in CO2 which is responsible for pCO2 values of more than 200 uatm below saturation. In addition, both measured and predicted values of pCO2 demonstrate large inter annual variability in the biological draw down of CO2 in the Ross Sea. While it is clear that remotely sensed ocean color data may be essential tool for monitoring inter annual variability in surface pCO2 throughout the world oceans, parameterization of mixed layer depths, upwelling and diapycnal diffusion from below the mixed layer and biological precipitation and dissolution of calcium carbonate are essential.

Dierdre A. Toole1, David A. Siegel1 and John W.H. Dacey2

Development of an upper-ocean dimethylsulfide (DMS) cycling model for the Sargasso Sea - a top-down approach

1 Institute for Computational Earth System Science, University of California, Santa Barbara, Santa Barbara, CA
2 Woods Hole Oceanographic Institution, Woods Hole, MA

Characterization and quantification of upper-ocean biogenic sulfur cycling is crucial for assessing its contribution to the flux of atmospheric sulfur and the oceanic carbon cycle. Recent studies have suggested that dimethylsulfoniopropionate (DMSP) and its environmentally important derivative, dimethylsulfide (DMS), account for up to 100% of total foodweb sulfur fluxes and approximately 5 - 15% of total foodweb carbon fluxes. Twice monthly vertical profiles of DMS and particulate and dissolved DMSP sampled in the Sargasso Sea at Hydrostation S between 1992 and 1994 are used to isolate relevant processes that contribute to upper ocean sulfur cycling. Regressions between in situ water column constituents and sulfur stocks produce poor results indicating that sulfur cycling is regulated by a variety of interdependent physical, chemical, and biological processes. DMS stocks correlate most strongly with chlorophyll, temperature, and optical properties however, suggesting that DMS cycling may be assessed via satellite retrieved quantities. This potential is explored using a 0-d model of biogenic sulfur driven by extensive time-series data sets collected from the Sargasso Sea at the Bermuda Atlantic Time-Series Study and Hydrostation S sites. Incorporated into this model are estimated sea to air fluxes and photo-oxidation rates, apparent quantum yields, and bacterial consumption rates measured in a variety of diverse environments. This suite of measurements allows for an assessment of the forcing mechanisms of DMS loss processes. Because of the complexity and lack of supporting measurements to constrain DMS production, a production based model would involve extensive, unresolved parameterizations and assumptions. Our goal is to use the results of this modeling effort to refine our understanding of DMS cycling and to serve as a basis for top-down satellite algorithms for upper ocean DMS stocks.

Daniela Turk1 (presenting poster by Bill Miller and Maurice Levasseur)

Canadian SOLAS Research Network

1 Canadian SOLAS Secretariat, Dalhousie University, Halifax, Nova Scotia, Canada, B3H 4J1

abstract This poster presents an overview of the Canadian SOLAS Research Network plans for 2001 through 2005. The scientific focus of the Canadian SOLAS research effort largely reflects its integration with the International SOLAS Program and its stated objective of addressing 'the key interactions among the marine biogeochemical system, the atmosphere and climate, and how this system affects and is affected by past and future climate and environmental changes'.

Specifically, the 15 individual Canadian projects, supporting 43 principal investigators, will support a science program with the following objectives.
1. Determine, during different seasons, the spatial distribution of trace gases production in major biogeochemical provinces of the northwest Atlantic and subarctic Pacific, and the impact of trace gases on the atmospheric chemical and physical properties and on climate.
2. Determine the influence of Fe on the production of trace gases in the SAP and their impact on the atmospheric chemical and physical properties and on climate.
3. Significantly increase our capacity to estimate trace gas emissions from whole oceanic basins using remote sensing.
4. Significantly increase our capacity to model ocean-atmosphere exchange over regional and seasonal scales.

Two major expeditions are planned: An Fe-addition in the subarctic Pacific where Fe limits primary production and an integrated study of a tagged patch of water in the northwest Atlantic during the spring bloom. Data collected to address this hypothesis include measurements of gas exchange dynamics, water column structure, trophic structure and physiological markers of the plankton community, rates of growth and elemental fluxes among key food-web components, remote sensing of ocean colour, and ultimate integration of field data into coupled ocean-atmosphere models.

Paul Tréguer1, Olivier Aumont2, Philippe Pondaven1, David M. Nelson3, Mark A. Brezinski4 and Nicolas Lemarchand1

3-D modeling of the primary and export production of carbon and biogenic silica in the Southern Ocean : assessing the real importance of the deep chlorophyll maximum layer

1 Université de Bretagne Occidentale, Institut Universitaire Européen de la Mer, UMR 6539, Brest-France
2 Université Pierre et Marie Curie, Institut Pierre Simon Laplace, Paris-France
3 Oregon State University, College of Oceanic and Atmospheric Sciences, Corvallis, OR 97331-5503, USA
4 University of California at Santa Barbara, California 93106-2050, USA

The inability of satellite sensors to detect frequently occurring sub-surface chlorophyll patches has been recently inferred to explain the discrepancy between the annual estimates of export production in the Southern Ocean, derived from satellite based productivity estimates and from outputs of inverse models. During numerous French, Australian, and United States spring and summer cruises large extended deep chlorophyll maximum (DCM), with concentrations up to 1.8 µg L-1, were detected in different sectors of the Southern Ocean south of 50°S. Non-limiting amounts of major nutrients (N, P) suggest that the growth of phytoplankton is limited by other factors, including top-down and/or bottom-up control. In these DCMs diatoms were generally significant contributors to carbon biomass and primary production. The 3D coupled physical - biogeochemical model (ORCA/PISCES) of the IPSL/LODyC is used to simulate the distribution of carbon and silica in the Southern Ocean. The modeled estimate for annual primary production south of 50°S is 4.8 GT C yr-1, the contribution of diatoms accounting for 1 GTC yr-1. The model simulates the formation of DCMs in the SIZas a result from co-limitation of light and Fe. The contribution of DCMs to the annual production of organic carbon and of biogenic silica in the Southern Ocean are important outputs of the model. On annual basis the DCMs contribute to 14% of the annual total. The interannual variability of the formation of DCMs over half a century will be also be presented.

J. D. Wiggert1, R. G. Murtugudde1 and J. R. Christian1

Annual ecosystem variability of the Indian Ocean: Results of a coupled bio-physical OGCM

1 Earth System Science Interdisciplinary Center, University of Maryland, College Park, MD, 20742-2465 jwiggert@essic.umd.edu

A fully coupled, 3-D bio-physical ocean general circulation model has been applied to the Indian Ocean (IO) basin. The biological portion of the model is a 9-component oceanic ecosystem, consisting of a large and small size class for phytoplankton, zooplankton and detritus, as well as three phytoplankton nutrients (nitrate, ammonium and iron). The ecosystem of the climatological solution presented here has been extensively validated. The NODC seasonal nitrate climatology and a monthly climatology of SeaWiFS ocean color provide validation over the full basin. The SeaWiFS climatology was created using the level 3, SMI (standard mapped image) monthly data (v3) from September 1997 through January 2002, which are routinely available from the Goddard DAAC (http://daac.gsfc.nasa.gov/). A more comprehensive validation of the model ecosystem in the Arabian Sea is facilitated by measurements of primary productivity, particulate organic nitrogen, zooplankton biomass, and particle flux obtained during the US JGOFS Arabian Sea Process Study. Comparing the solution to the diverse validation data applied to characterize the ecosystem's behavior in the Arabian Sea demonstrates that all biogeochemical processes are reasonably represented. Furthermore, the model's overall success at simulating the considerable temporal and spatial biogeochemical variability characteristic of the Arabian Sea engenders confidence in the model solution in regions of the IO that are observation deficient. Finally, we capitalize on this IO basin scale bio-physical model with explicit iron biogeochemistry to provide new insights into biogeochemical processes associated with several annually recurringdynamical features (e.g., the Somali Current system) and the seasonal evolution of the basinwide distribution of iron limited phytoplankton growth.

James A. Yoder1 and Maureen A. Kennelly1

Seasonal and ENSO Variability in Global Ocean Phytoplankton Chlorophyll

1 Graduate School of Oceanography, University of Rhode Island, Narragansett, RI 02882

Seasonal changes in phytoplankton biomass and productivity are very important components of the total variability associated with ocean biological and biogeochemical processes. Seasonal changes in phytoplankton biomass and productivity are generally related to incident solar irradiance, upper ocean mixing and stratification and other processes that affect the ocean's light and nutrient environment. Satellite ocean color sensors now provide data sets to assess seasonal and other sources of phytoplankton variability on global scales. Imagery from two satellite ocean color missions (OCTS: 1996-1997 and SeaWiFS: 1998-2001) has been prepared for LAS distribution at 1 degree x 1 degree spatial resolution and 8-day temporal resolution. To quanitify the major seasonal (as well as the 1998 ENSO) signals in phytoplankton biomass between 50°S and 50°N we used empirical orthogonal function (EOF) analysis on the 4-year SeaWiFS time series of global chlorophyll a. We then did a second ananlysis of the SeaWiFS data set to quanitify summer patterns at higher latitudes. Among the important effects we resolved are a 6-month phase shift in maximum chlorophyll a concentrations between subtropical (winter peaks) and subpolar (spring-summer peaks) waters, greater seasonal range at high latitudes in the Atlantic compared to the Pacific, an interesting phasing between spring and fall biomass peaks at high latitudesin both hemispheres, and the effects of the 1998 ENSO cycle in the tropics. Our EOF results show that dominant seasonal and ENSO effects are captured in the first 6 of a possible 184 modes, which explain 68% of the total temporal variability associated with the global mean phytonplankton chlorophyll pattern in our smoothed data set. The results also show that the time (seasonal)/space (zonal) patterns between the ocean basins and between the hemispheres are similar, albeit with some key differences. Finally, the dominant global patterns are consistent with the results of ocean models of seasonal dynamics based on seasonal changes to the heating and cooling (stratification/de-stratificiation) cycles of the upper ocean.



1 author address
2 author address